Datasets:

DSGym
/

DSGym-Tasks

Error code:   DatasetGenerationError
Exception:    ArrowTypeError
Message:      ("Expected bytes, got a 'list' object", 'Conversion failed for column question with type object')
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 160, in _generate_tables
                  pa_table = paj.read_json(
                             ^^^^^^^^^^^^^^
                File "pyarrow/_json.pyx", line 342, in pyarrow._json.read_json
                File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
                File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
              pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to array in row 0
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1869, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 190, in _generate_tables
                  pa_table = pa.Table.from_pandas(df, preserve_index=False)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "pyarrow/table.pxi", line 4795, in pyarrow.lib.Table.from_pandas
                File "/usr/local/lib/python3.12/site-packages/pyarrow/pandas_compat.py", line 637, in dataframe_to_arrays
                  arrays = [convert_column(c, f)
                            ^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pyarrow/pandas_compat.py", line 625, in convert_column
                  raise e
                File "/usr/local/lib/python3.12/site-packages/pyarrow/pandas_compat.py", line 619, in convert_column
                  result = pa.array(col, type=type_, from_pandas=True, safe=safe)
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "pyarrow/array.pxi", line 365, in pyarrow.lib.array
                File "pyarrow/array.pxi", line 91, in pyarrow.lib._ndarray_to_array
                File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
              pyarrow.lib.ArrowTypeError: ("Expected bytes, got a 'list' object", 'Conversion failed for column question with type object')
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1347, in compute_config_parquet_and_info_response
                  parquet_operations = convert_to_parquet(builder)
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 980, in convert_to_parquet
                  builder.download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 884, in download_and_prepare
                  self._download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 947, in _download_and_prepare
                  self._prepare_split(split_generator, **prepare_split_kwargs)
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1736, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1919, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

context string	question string	answer string	data list	metadata dict
No context is necessary for this task.	Calculate the mean fare paid by the passengers.	[['mean_fare', '34.65']]	[ "/data/daeval/test_ave.csv" ]	{ "constraints": "Calculate the mean fare using Python's built-in statistics module or appropriate statistical method in pandas. Rounding off the answer to two decimal places.", "format": "@mean_fare[mean_fare_value] where \"mean_fare_value\" is a floating-point number rounded to two decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 0 }
No context is necessary for this task.	Generate a new feature called "FamilySize" by summing the "SibSp" and "Parch" columns. Then, calculate the Pearson correlation coefficient (r) between the "FamilySize" and "Fare" columns.	[['correlation_coefficient', '0.21']]	[ "/data/daeval/test_ave.csv" ]	{ "constraints": "Create a new column 'FamilySize' that is the sum of 'SibSp' and 'Parch' for each row.\nCalculate the Pearson correlation coefficient between 'FamilySize' and 'Fare'\nDo not perform any further data cleaning or preprocessing steps before calculating the correlation.", "format": "@correlation_coefficient[r_value]\nwhere \"r_value\" is the Pearson correlation coefficient between 'FamilySize' and 'Fare', a number between -1 and 1, rounded to two decimal places.", "keywords": [ "Feature Engineering", "Correlation Analysis" ], "level": "medium", "original_id": 5 }
No context is necessary for this task.	Create a new column called "AgeGroup" that categorizes the passengers into four age groups: 'Child' (0-12 years old), 'Teenager' (13-19 years old), 'Adult' (20-59 years old), and 'Elderly' (60 years old and above). Then, calculate the mean fare for each age group.	[['mean_fare_elderly', '43.47'], ['mean_fare_teenager', '31.98'], ['mean_fare_child', '31.09'], ['mean_fare_adult', '35.17']]	[ "/data/daeval/test_ave.csv" ]	{ "constraints": "Make sure to round the mean fare of each group to 2 decimal places.", "format": "@mean_fare_child[mean_fare], @mean_fare_teenager[mean_fare], @mean_fare_adult[mean_fare], @mean_fare_elderly[mean_fare], where \"mean_fare\" is a float number rounded to 2 decimal places.", "keywords": [ "Feature Engineering", "Summary Statistics" ], "level": "medium", "original_id": 6 }
No context is necessary for this task.	Apply the linear regression algorithm from the sklearn library to predict whether a passenger survived or not based on the features 'Pclass', 'Sex', 'Age', 'SibSp', 'Parch', 'Fare', and 'Embarked'. Encode 'Sex' and 'Embarked' to numerical values before applying the model. Split the dataset into a training set (80%) and a testing set (20%), train the model on the training set, and evaluate its performance on the testing set using the accuracy score. Ensure that the train_test_split function's random_state parameter is set to 42 for consistency.	[['prediction_accuracy', '0.78']]	[ "/data/daeval/test_ave.csv" ]	{ "constraints": "Use one-hot encoding for the 'Sex' and 'Embarked' features. Use the \"linear regression\" model provided by the sklearn library in Python.", "format": "@prediction_accuracy[accuracy], where \"accuracy\" is a float number rounded to 2 decimal places and has a range of 0.0 to 1.0.", "keywords": [ "Machine Learning" ], "level": "hard", "original_id": 7 }
No context is necessary for this task.	Calculate the mean value of the "Close Price" column.	[['mean_close_price', '570.68']]	[ "/data/daeval/GODREJIND.csv" ]	{ "constraints": "Use the built-in Python (numpy or pandas) to calculate the mean. Do not use any pre-built packages or libraries for mean calculation other than numpy or pandas. The calculation should be done on the whole \"Close Price\" column. Values in this column should not be rounded or changed in any way before the calculation.", "format": "@mean_close_price[mean_value], where \"mean_value\" is a float number rounded to two decimal places. This value should be between the highest and lowest \"Close Price\" given in the dataset.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 9 }
No context is necessary for this task.	Create a new feature called "Price Range" which represents the difference between the "High Price" and "Low Price" for each row. Calculate the mean, median, and standard deviation of this new feature.	[['price_range_mean', '16.65'], ['price_range_std_dev', '6.72'], ['price_range_median', '15.67']]	[ "/data/daeval/GODREJIND.csv" ]	{ "constraints": "Make sure to use the correct columns for calculating the \"Price Range\". All calculations should be performed up to two decimal places.", "format": "@price_range_mean[mean]: The mean should be a single real number rounded to two decimal places. @price_range_median[median]: The median should be a single real number rounded to two decimal places. @price_range_std_dev[std_dev]: The standard deviation should be a single real number rounded to two decimal places.", "keywords": [ "Feature Engineering", "Summary Statistics" ], "level": "medium", "original_id": 14 }
No context is necessary for this task.	Calculate the mean and standard deviation of the "Mar.2019" column.	[['mean_mar_2019', '171.44'], ['sd_mar_2019', '188.25']]	[ "/data/daeval/unemployement_industry.csv" ]	{ "constraints": "Outliers are to be pruned via the interquartile range method before calculating the mean and standard deviation. Handle missing values by using listwise deletion method. Report the measures rounded to two decimal places.", "format": "@mean_mar_2019[mean] @sd_mar_2019[sd], where \"mean\" and \"sd\" are numbers, rounded to two decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 18 }
No context is necessary for this task.	Apply machine learning techniques to predict the employment level in March 2020 based on the data from March 2019. Split the dataset into a 70-30 split for training and testing sets, train a simple linear regression model on the training set, and evaluate its performance on the testing set using Mean Squared Error as the evaluation metric.	[['Mean_Squared_Error', '11439.60']]	[ "/data/daeval/unemployement_industry.csv" ]	{ "constraints": "Split the dataset with a 70-30 split for training and testing sets with a random seed of 42. Use a simple linear regression model for training and evaluate the model's performance by calculating the Mean Squared Error.", "format": "@Mean_Squared_Error[MSE], where \"MSE\" is a number rounded to two decimal places.", "keywords": [ "Machine Learning", "Summary Statistics" ], "level": "hard", "original_id": 23 }
No context is necessary for this task.	Calculate the mean age of the individuals in the dataset.	[['mean_age', '39.21']]	[ "/data/daeval/insurance.csv" ]	{ "constraints": "Ignore rows with missing values in the age column. Use Python's built-in function to calculate the mean.", "format": "@mean_age[value] where \"value\" is a number between 0 and 100, rounded to two decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 24 }
No context is necessary for this task.	Check if the distribution of BMI values in the dataset follows a normal distribution.	[['bmi_distribution', 'normal']]	[ "/data/daeval/insurance.csv" ]	{ "constraints": "Consider the distribution as normal if the absolute value of skewness is less than 0.5. Calculate skewness using Python's built-in function.", "format": "@bmi_distribution[status] where \"status\" is a string that can either be \"normal\" or \"not_normal\", based on the conditions specified in the constraints.", "keywords": [ "Distribution Analysis" ], "level": "easy", "original_id": 25 }
No context is necessary for this task.	Calculate the correlation coefficient between the charges incurred by individuals and the number of children they have.	[['correlation_coefficient', '0.07']]	[ "/data/daeval/insurance.csv" ]	{ "constraints": "Ignore rows with missing values in charges and children columns. Calculate the Pearson correlation coefficient.", "format": "@correlation_coefficient[value] where \"value\" is a number between -1 and 1, rounded to two decimal places.", "keywords": [ "Correlation Analysis" ], "level": "easy", "original_id": 26 }
No context is necessary for this task.	Identify the outliers in the charges incurred by individuals using the Z-score method.	[['median_charges_outliers', '40974.16'], ['mean_charges_outliers', '42103.95'], ['total_outliers', '139']]	[ "/data/daeval/insurance.csv" ]	{ "constraints": "Apply the Z-score method for outlier detection using the 1.5xIQR rule. Consider any value that falls below Q1 - 1.5 * IQR or above Q3 + 1.5 * IQR as an outlier. Report the total number of outliers, and the mean and median charges of these identified outliers.", "format": "@total_outliers[total_outliers] \n@mean_charges_outliers[mean_charges_outliers] \n@median_charges_outliers[median_charges_outliers] \nwhere \"total_outliers\" is an integer, \"mean_charges_outliers\" and \"median_charges_outliers\" are floating-point numbers rounded to two decimal places.", "keywords": [ "Outlier Detection" ], "level": "medium", "original_id": 27 }
No context is necessary for this task.	Create a linear regression machine learning model using the Scikit-learn library to predict the medical charges based on the age and BMI of individuals. Evaluate the performance of the model using the Root Mean Square Error (RMSE) evaluation metric only.	[['model_rmse', '11464.74']]	[ "/data/daeval/insurance.csv" ]	{ "constraints": "Use the linear regression model available in the Scikit-Learn library. Split the data into training and testing sets with 80% of the data used for training and 20% used for testing. Use a random state of 42 for the split. The predictor variables are 'age' and 'bmi', and the target variable is 'charges'. Implement RMSE for the model evaluation. Ignore any row with missing values present in these three columns for this analysis.", "format": "@model_rmse[RMSE_value], where RMSE value is a positive number rounded to two decimal places.", "keywords": [ "Machine Learning", "Feature Engineering" ], "level": "hard", "original_id": 30 }
No context is necessary for this task.	Calculate the mean and standard deviation of the "importance.score" column.	[['importance_score_std', '0.01'], ['importance_score_mean', '0.00']]	[ "/data/daeval/imp.score.ldlr.metabolome.csv" ]	{ "constraints": "Calculate the mean and standard deviation to two decimal places for the \"importance.score\" column. Ignore any null or missing values in the calculations. The calculations are to be done using standard statistical methods without applying any transformations or filters to the data.", "format": "@importance_score_mean[mean] @importance_score_std[std_dev] where \"mean\" and \"std_dev\" are non-negative numbers rounded to two decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 32 }
No context is necessary for this task.	Is there a correlation between the "row retention time" and "importance.score" columns?	[['p_value', '0.4058'], ['relationship_type', 'none'], ['correlation_coefficient', '-0.04']]	[ "/data/daeval/imp.score.ldlr.metabolome.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between \"row retention time\" and \"importance.score\". Assess the significance of the correlation using a two-tailed test with a significance level (alpha) of 0.05. Report the p-value associated with the correlation test. Consider the relationship to be linear if the p-value is less than 0.05 and the absolute value of r is greater than or equal to 0.5. Consider the relationship to be nonlinear if the p-value is less than 0.05 and the absolute value of r is less than 0.5. If the p-value is greater than or equal to 0.05, report that there is no significant correlation. Ignore any null or missing values in performing the correlation test.", "format": "@correlation_coefficient[r_value] @p_value[p_value] @relationship_type[relationship_type] where \"r_value\" is a number between -1 and 1, rounded to two decimal places. \"p_value\" is a number between 0 and 1, rounded to four decimal places. \"relationship_type\" is a string that can either be \"linear\", \"nonlinear\", or \"none\" based on the conditions specified in the constraints.", "keywords": [ "Correlation Analysis" ], "level": "medium", "original_id": 34 }
No context is necessary for this task.	Identify and remove any outliers in the "row retention time" column using the Z-score method with a Z-score threshold of 3. Provide the number of removed outliers.	[['removed_outliers_count', '0']]	[ "/data/daeval/imp.score.ldlr.metabolome.csv" ]	{ "constraints": "Use the Z-score method to identify outliers in the \"row retention time\" column. Any data point with a Z-score greater than 3 or less than -3 is considered an outlier and should be removed.", "format": "@removed_outliers_count[count] where \"count\" is a non-negative integer indicating the count of removed outliers.", "keywords": [ "Outlier Detection", "Comprehensive Data Preprocessing" ], "level": "medium", "original_id": 35 }
No context is necessary for this task.	What is the mean number of cases recorded across all countries and years?	[['mean_cases', '2081990']]	[ "/data/daeval/estimated_numbers.csv" ]	{ "constraints": "Calculate the mean of the column 'No. of cases'. Convert the data type of 'No. of cases' column from Object (string) to Int64 before performing calculations. Ignore those records where 'No. of cases' column value is Null or empty.", "format": "@mean_cases[mean_value] where \"mean_value\" is a positive integer.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 55 }
No context is necessary for this task.	Which country has the highest number of deaths recorded in a single year?	[['max_deaths_country', 'Nigeria'], ['max_deaths_year', '2010']]	[ "/data/daeval/estimated_numbers.csv" ]	{ "constraints": "Calculate the maximum value in the 'No. of deaths' column. Convert the data type of 'No. of deaths' column from Object (string) to Int64 before performing calculations. Ignore those records where 'No. of deaths' column value is Null or empty. Identify the corresponding country and year for the highest number of deaths.", "format": "@max_deaths_country[country_name] @max_deaths_year[year] where \"country_name\" is a string indicating the name of the country and \"year\" is an integer indicating the year in which the maximum deaths occurred.", "keywords": [ "Distribution Analysis", "Summary Statistics" ], "level": "easy", "original_id": 56 }
No context is necessary for this task.	Is there a correlation between the number of cases and the number of deaths recorded?	[['correlation_coefficient', '0.97']]	[ "/data/daeval/estimated_numbers.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between number of cases and number of deaths. Convert the data types of 'No. of cases' and 'No. of deaths' column from Object (String) to Int64 before performing calculations. Do this for complete data rather than specific country or year.", "format": "@correlation_coefficient[r_value] where \"r_value\" is a number between -1 and 1, rounded to two decimal places.", "keywords": [ "Correlation Analysis" ], "level": "easy", "original_id": 57 }
No context is necessary for this task.	What is the percentage of missing values in the "No. of cases_min" column? How does this percentage compare to the percentage of missing values in the "No. of deaths_max" column?	[['percentage_cases_min', '36.45'], ['percentage_deaths_max', '38.79']]	[ "/data/daeval/estimated_numbers.csv" ]	{ "constraints": "Calculate the percentage of missing values for both \"No. of cases_min\" and \"No. of deaths_max\" column. Report the exact percentage values.", "format": "@percentage_cases_min[percentage], @percentage_deaths_max[percentage] where \"percentage\" is a number between 0 and 100, rounded to two decimal places.", "keywords": [ "Comprehensive Data Preprocessing", "Summary Statistics" ], "level": "easy", "original_id": 58 }
No context is necessary for this task.	Among the countries in the "Americas" region, which country has the highest average number of cases recorded over the years?	[['country_name', 'Congo']]	[ "/data/daeval/estimated_numbers.csv" ]	{ "constraints": "Calculate the average of \"No. of cases\" for each country in the \"Americas\" region and report the country with the highest average number of cases. Count only complete years, i.e., exclude years with missing data.", "format": "@country_name[country] where \"country\" is a string representing the name of the country with the highest average number of cases.", "keywords": [ "Distribution Analysis", "Summary Statistics", "Feature Engineering" ], "level": "easy", "original_id": 59 }
No context is necessary for this task.	Calculate the mean and standard deviation of the wage column.	[['std_wage', '4.66'], ['mean_wage', '6.31']]	[ "/data/daeval/beauty and the labor market.csv" ]	{ "constraints": "The mean and standard deviation of the wage should be calculated using pandas' `mean()` and `std()` methods respectively. Do not apply any transformations, filtering or alteration to the wage data.", "format": "@mean_wage[mean_value] @std_wage[std_value] where \"mean_value\" and \"std_value\" are numbers with up to two decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 64 }
No context is necessary for this task.	Perform feature engineering by creating a new feature called "experience_score" that is calculated by multiplying the exper column with the looks column. Then, calculate the Pearson correlation coefficient between the "experience_score" feature and the wage column.	[['correlation', '0.252']]	[ "/data/daeval/beauty and the labor market.csv" ]	{ "constraints": "Create \"experience_score\" by directly multiplying values of exper and looks column. Calculate Pearson correlation coefficient between the new feature \"experience_score\" and wage. Correlation should be calculated up to three decimal places.", "format": "@correlation[correlation] where \"correlation\" is a number representing the correlation coefficient, rounded to three decimal places.", "keywords": [ "Feature Engineering", "Correlation Analysis" ], "level": "medium", "original_id": 69 }
No context is necessary for this task.	Perform machine learning by training a linear regression model to predict the wage based on the features exper, looks, union, goodhlth, black, female, married, south, bigcity, smllcity, service, and educ. Use the Root Mean Squared Error (RMSE) for evaluating the model's performance.	[['RMSE', '3.63']]	[ "/data/daeval/beauty and the labor market.csv" ]	{ "constraints": "Perform the machine learning task using the sklearn library's LinearRegression() function. Split the dataset into a 70% training set and a 30% test set. Set the random seed to 42 for reproducibility of the results. Calculate the RMSE on the test set.", "format": "@RMSE[RMSE_value] where \"RMSE_value\" is a number rounded to two decimal places.", "keywords": [ "Machine Learning", "Summary Statistics" ], "level": "hard", "original_id": 70 }
No context is necessary for this task.	Calculate the mean and standard deviation of the "Volume" column.	[['mean_volume', '22607406.19'], ['std_dev_volume', '8254791.71']]	[ "/data/daeval/microsoft.csv" ]	{ "constraints": "Use the built-in functions in Python's pandas library for computation. Round the result to 2 decimal places.", "format": "@mean_volume[mean_volume]\\n@std_dev_volume[std_dev_volume]\\nwhere \"mean_volume\" and \"std_dev_volume\" are floating-point numbers rounded to 2 decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 71 }
No context is necessary for this task.	Check if the "Close" column adheres to a normal distribution.	[['normality_test_result', 'Non-normal']]	[ "/data/daeval/microsoft.csv" ]	{ "constraints": "Use the Shapiro-Wilk test to assess the normality of the \"Close\" column. If the p-value is less than 0.05, consider the data to be non-normally distributed. Otherwise, consider it to be normally distributed.", "format": "@normality_test_result[normality_test_result]\\nwhere \"normality_test_result\" a string that is either \"Normal\" or \"Non-normal\" based on the p-value from the Shapiro-Wilk test.", "keywords": [ "Distribution Analysis" ], "level": "easy", "original_id": 72 }
No context is necessary for this task.	Calculate the correlation coefficient between the "High" and "Low" columns.	[['correlation_coefficient', '1.0']]	[ "/data/daeval/microsoft.csv" ]	{ "constraints": "Use the Pearson correlation coefficient for computation. Round the result to 2 decimal places.", "format": "@correlation_coefficient[correlation_coefficient]\\nwhere \"correlation_coefficient\" is a floating-point number rounded to 2 decimal places.", "keywords": [ "Correlation Analysis" ], "level": "easy", "original_id": 73 }
No context is necessary for this task.	Create a new column called "Daily Return" that calculates the percentage change in the "Close" price from the previous day. Calculate the mean and standard deviation of the "Daily Return" column.	[['daily_return_std', '0.94'], ['daily_return_mean', '-0.14']]	[ "/data/daeval/microsoft.csv" ]	{ "constraints": "Calculate \"Daily Return\" as ((Close price of today - Close price of previous day) / Close price of previous day) * 100. Calculate mean and standard deviation to two decimal places.", "format": "@daily_return_mean[mean], @daily_return_std[std] where \"mean\" and \"std\" are the mean and standard deviation of the \"Daily Return\" column, respectively, rounded to two decimal places.", "keywords": [ "Feature Engineering", "Summary Statistics" ], "level": "medium", "original_id": 75 }
No context is necessary for this task.	Perform comprehensive data preprocessing on the "Date" column to extract the month and year information. Calculate the average closing price for each month and year combination. Return the month and year combination which has the highest average closing price.	[['Highest_Monthly_Average_Close_Price', '1, 2018, 88.32']]	[ "/data/daeval/microsoft.csv" ]	{ "constraints": "Set the start of each month to be the first day of the month and the end of the month to be the last day of the month.\nThe calculation of the average closing price should be done using the arithmetic mean. \nFor ties, return the most recent month and year combination.", "format": "@Highest_Monthly_Average_Close_Price[month, year, average_close_price]\nwhere \"month\" is an integer between 1 and 12, \"year\" is an integer, and \"average_close_price\" is a float rounded to two decimal places.", "keywords": [ "Comprehensive Data Preprocessing", "Summary Statistics" ], "level": "hard", "original_id": 77 }
No context is necessary for this task.	Calculate the correlation coefficient between ApplicantIncome and LoanAmount.	[['correlation_coefficient', '0.49']]	[ "/data/daeval/test_Y3wMUE5_7gLdaTN.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient to assess the strength and direction of linear relationship between ApplicantIncome and LoanAmount. Ignore the rows with missing values for either of the two columns. Round the correlation coefficient to two decimal places.", "format": "@correlation_coefficient[corr_coeff] where \"corr_coeff\" is a number between -1 and 1, rounded to two decimal places and represents the Pearson correlation coefficient between ApplicantIncome and LoanAmount.", "keywords": [ "Correlation Analysis" ], "level": "medium", "original_id": 105 }
No context is necessary for this task.	Generate a new feature called "TotalIncome" by adding the ApplicantIncome and CoapplicantIncome columns. Calculate the mean and standard deviation of the TotalIncome column.	[['mean_total_income', '6375.18'], ['std_dev_total_income', '5199.42']]	[ "/data/daeval/test_Y3wMUE5_7gLdaTN.csv" ]	{ "constraints": "Calculate the mean and standard deviation using the Panda's DataFrame mean() and std() functions distinctively. Round the results to two decimal places.", "format": "@mean_total_income[mean] @std_dev_total_income[std_dev] where \"mean\" is a float number that represents the mean value of the TotalIncome column rounded to two decimal places, and \"std_dev\" is a float number that represents the standard deviation of the TotalIncome column also rounded to two decimal places.", "keywords": [ "Feature Engineering", "Summary Statistics" ], "level": "medium", "original_id": 108 }
No context is necessary for this task.	Explore the distribution of the LoanAmount column based on different values of the Education column. Determine if there is a significant difference in the loan amount between individuals with different educational backgrounds.	[['graduate_mean_loan', '141.36'], ['not_graduate_mean_loan', '118.57'], ['significance', 'significant']]	[ "/data/daeval/test_Y3wMUE5_7gLdaTN.csv" ]	{ "constraints": "Calculate the mean of LoanAmount for individuals with a 'Graduate' educational background and individuals with a 'Not Graduate' educational background separately. Test if there is a significant difference between these two groups using a t-test with a significance level (alpha) of 0.05. If the p-value is less than 0.05, report there is a significant difference, else report there is no significant difference.", "format": "@graduate_mean_loan[mean], @not_graduate_mean_loan[mean], @significance[significant/no significant] where \"mean\" is a number (float), rounded to two decimal places. \"significant\" or \"no significant\" signifies if there is a significant difference between two groups under the significance level 0.05.", "keywords": [ "Distribution Analysis", "Feature Engineering" ], "level": "hard", "original_id": 109 }
No context is necessary for this task.	Which country has the highest happiness score?	[['country_with_highest_score', 'Switzerland']]	[ "/data/daeval/2015.csv" ]	{ "constraints": "Find the country with the highest happiness score in the dataset. If two or more countries have the same highest happiness score, return all of them.", "format": "@country_with_highest_score[country_name]", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 114 }
No context is necessary for this task.	Which variable has the strongest correlation with the happiness scores among countries? Is this correlation positive or negative?	[['correlation_type', 'negative'], ['strongest_correlation_variable', 'Happiness Rank']]	[ "/data/daeval/2015.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient (r) between the happiness score and all other numerical variables in the dataset. The variable which has the highest magnitude of r (ignoring the sign) is the one with the strongest correlation.", "format": "@strongest_correlation_variable[variable_name] where \"variable_name\": the column name of the variable with the strongest correlation. @correlation_type[positive/negative] where \"positive/negative\": if the correlation is positive or negative based on the sign of the correlation coefficient.", "keywords": [ "Correlation Analysis" ], "level": "medium", "original_id": 117 }
No context is necessary for this task.	Is there a linear relationship between the GDP per capita and the life expectancy score in the dataset? Conduct linear regression and use the resulting coefficient of determination (R-squared) to evaluate the model's goodness of fit.	[['coefficient_determination', '0.67'], ['model_fit', 'poor fit']]	[ "/data/daeval/2015.csv" ]	{ "constraints": "Calculate the coefficient of determination (R-squared) for the given relationship. If R-squared is equal to or greater than 0.7, consider the model a good fit. Else, consider it a poor fit.", "format": "@coefficient_determination[R_square], @model_fit[model_fit], where \"R_square\" is the value of the coefficient of determination rounded to two decimal places and \"model_fit\" is a string that is either \"good fit\" or \"poor fit\" based on the calculated R-squared value.", "keywords": [ "Correlation Analysis", "Machine Learning" ], "level": "hard", "original_id": 118 }
No context is necessary for this task.	Which country has the highest average number of daily vaccinations per million people?	[['country_with_highest_average_daily_vaccinations', 'Gibraltar']]	[ "/data/daeval/country_vaccinations.csv" ]	{ "constraints": "{\nBased on the current available data without null values in the column of daily vaccinations per million people.\nNo tie of the maximum value is allowed. In case of a tie, consider the country with the first appeared maximum value.\n}", "format": "{\n@country_with_highest_average_daily_vaccinations[country_name]\nWhere \"country_name\" is a string.", "keywords": [ "Summary Statistics", "Distribution Analysis" ], "level": "easy", "original_id": 123 }
No context is necessary for this task.	Can we predict the number of people fully vaccinated per hundred people based on the total number of vaccinations administered and the number of people vaccinated per hundred people?	[['significant_predictor', 'yes,yes'], ['r_squared', '0.6059']]	[ "/data/daeval/country_vaccinations.csv" ]	{ "constraints": "{\nPerform a multiple linear regression analysis using the total number of vaccinations administered and the number of people vaccinated per hundred people as predictors.\nThe dependent variable is the number of people fully vaccinated per hundred people.\nOnly consider data entries without null values in the three mentioned columns.\nUse a significance level (alpha) of 0.05 for the predictors.\nConsider the predictors to be significant if the p-value is less than 0.05.\nCalculate the R-squared value of the model.\n}", "format": "{\n@significant_predictor[predictor_1,predictor_2]\n@r_squared[r_squared_value]\nWhere \"predictor_1,predictor_2\" can be \"yes,yes\", \"yes,no\", \"no,yes\", or \"no,no\" based on the p-values of the predictors.\nWhere \"r_squared_value\" is a number between 0 and 1, rounded to four decimal places.", "keywords": [ "Correlation Analysis", "Machine Learning" ], "level": "hard", "original_id": 125 }
No context is necessary for this task.	Identify and count the number of outliers in the fare paid by passengers using the Z-score method.	[['outlier_count', '20']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "{\nCalculate the Z-score for each fare using the mean and standard deviation of the fare data.\nDetermine an outlier to be any fare with a Z-score greater than 3 or less than -3.\n}", "format": "{\n@outlier_count[count]\nwhere \"count\" is a positive integer or zero representing the total number of outliers found.", "keywords": [ "Outlier Detection" ], "level": "medium", "original_id": 132 }
No context is necessary for this task.	Perform comprehensive data preprocessing for the dataset by handling missing values in the age and cabin columns. Use the deletion strategy for the missing values in the cabin column and imputation strategy for the missing values in the age column.	[['median_age', '36.0'], ['row_count', '204']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "{\nFor the deletion strategy in the cabin column, remove any row that has a missing value in the cabin column.\nFor the imputation strategy in the age column, replace the missing values with the median age of all passengers.\nReport on the new total number of rows after deletion and the median age used for imputation.\n}", "format": "{\n@row_count[count]\n@median_age[value]\nwhere \"count\" is a positive integer representing the total number of rows after deletion.\nwhere \"value\" is a positive real number representing the median age used for imputation, rounded to one decimal place.", "keywords": [ "Comprehensive Data Preprocessing" ], "level": "hard", "original_id": 133 }
No context is necessary for this task.	Perform distribution analysis on the fare paid by passengers for each passenger class separately. Use the Shapiro-Wilk Test for normality. For each passenger class, the null hypothesis is that the fare follows a normal distribution.	[['p_value_class_1', '0.0000'], ['p_value_class_3', '0.0000'], ['p_value_class_2', '0.0000']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "The Shapiro-Wilk Test should be conducted separately for each passenger class. Use a significance level (alpha) of 0.05. If the p-value is less than 0.05, reject the null hypothesis.", "format": "@p_value_class_1[p_value_1], @p_value_class_2[p_value_2], @p_value_class_3[p_value_3] where 'p_value_1', 'p_value_2', and 'p_value_3' are the p-values of the Shapiro-Wilk Test for the 1st, 2nd and 3rd class passengers respectively, rounded to four decimal places.", "keywords": [ "Distribution Analysis", "Summary Statistics" ], "level": "medium", "original_id": 136 }
No context is necessary for this task.	Perform feature engineering by creating a new binary feature called "IsAlone" that indicates whether a passenger is traveling alone or with family. Use the "SibSp" and "Parch" columns to determine if a passenger has any accompanying family members. Then, train a logistic regression machine learning model using the new feature and the Survival rate as the output variable.	[['model_score', '0.61']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "The logistic regression model should be implemented with scikit-learn’s LogisticRegression with default parameters. Use the 'IsAlone' feature and 'Survived' as the output variable. The model should be trained using a 70:30 train-test split, balancing the class weights. Use a random seed of 42 for reproducibility.", "format": "@model_score[model_accuracy] where 'model_accuracy' is a number between 0 and 1, rounded to 2 decimal places, representing the accuracy of the model on the test set.", "keywords": [ "Feature Engineering", "Machine Learning" ], "level": "hard", "original_id": 137 }
No context is necessary for this task.	Question 2: Are the percentage of votes received by the Democratic party in a particular county normally distributed?	[['normality_status', 'not normal']]	[ "/data/daeval/election2016.csv" ]	{ "constraints": "{Test whether the 'per_dem' column follows a normal distribution using the Shapiro-Wilk test for normality. Set the significance level (alpha) at 0.05. If p-value is less than 0.05, reject the null hypothesis and report that the data is not normally distributed. If p-value is greater than or equal to 0.05, fail to reject the null hypothesis and report that the data is normally distributed.}", "format": "{@normality_status[status] where \"status\" is a string that could either be \"normal\" when the data follows normal distribution or \"not normal\" when the data does not follow a normal distribution.", "keywords": [ "Distribution Analysis" ], "level": "easy", "original_id": 139 }
No context is necessary for this task.	Question 2: Is there a relationship between the difference in votes received by the Democratic and Republican parties and their percentage point difference?	[['relationship_type', 'none'], ['correlation_coefficient', '0.02'], ['p_value', '0.1704']]	[ "/data/daeval/election2016.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between the difference in votes and the percentage point difference. Assess the significance of the correlation using a two-tailed test with a significance level (alpha) of 0.05. Report the p-value associated with the correlation test. Consider the relationship to be linear if the p-value is less than 0.05 and the absolute value of r is greater than or equal to 0.5. Consider the relationship to be nonlinear if the p-value is less than 0.05 and the absolute value of r is less than 0.5. If the p-value is greater than or equal to 0.05, report that there is no significant correlation.", "format": "@correlation_coefficient[r_value] @p_value[p_value] @relationship_type[relationship_type] where r_value is a number between -1 and 1, rounded to two decimal places. Where p_value is a number between 0 and 1, rounded to four decimal places. Where relationship_type is a string that can either be \"linear\", \"nonlinear\", or \"none\" based on the conditions specified in the constraints.", "keywords": [ "Correlation Analysis" ], "level": "hard", "original_id": 142 }
No context is necessary for this task.	Determine the skewness of the fares paid by the passengers on the Titanic.	[['fare_skewness', '4.79']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "For the calculation of skewness, use the pandas DataFrame method skew(). No other method should be employed for calculation.", "format": "@fare_skewness[fare_skew_value] \nwhere \"fare_skew_value\" is a floating-point number rounded to 2 decimal places.", "keywords": [ "Distribution Analysis" ], "level": "easy", "original_id": 174 }
No context is necessary for this task.	Calculate the median age of male passengers who survived and paid a fare greater than the average fare. Calulate only the ages that are not null.	[['median_age', '31.50']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "{\nAll null values in the \"Age\" column are not considered in the calculation.\nThe passengers considered for this question should meet all the following conditions: they are male; they survived; their fare is greater than the average fare.\n}", "format": "{\n@median_age[median_age]\nwhere \"median_age\" is a number rounded to two decimal places.", "keywords": [ "Summary Statistics", "Correlation Analysis" ], "level": "medium", "original_id": 176 }
No context is necessary for this task.	Perform comprehensive data preprocessing on the dataset. Handle missing values in the "Embarked" column by imputing them with the mode value. Normalize the "Fare" column using Min-Max scaling. Encode the categorical variable "Sex" using Label Encoding, where "male" is coded as 1 and "female" as 0. Calculate the number of each label after processing "Sex" and the minimum, maximum and mean of "Fare" after scaling.	[['sex_encoded_count', '314, 577'], ['fare_after_scaling', '0.00, 1.00, 0.0629']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "{\n\"Embarked\" missing values should be filled with the mode value.\n\"Fare\" should be normalized using Min-Max scaling, where Min is the minimum value of \"Fare\" before scaling and Max is the maximum.\n\"Sex\" should be encoded using Label Encoding, where \"male\" is 1 and \"female\" is 0.\nCaculate the count of each label of \"Sex\" after encoding and the min, max and mean of \"Fare\" values after scaling.\n}", "format": "{\n@sex_encoded_count[label_0_count, label_1_count]\n@fare_after_scaling[min_fare, max_fare, mean_fare]\nwhere \"label_0_count\" and \"label_1_count\" are numbers indicating the count of 0 and 1 in the encoded \"Sex\" column, respectively.\nwhere \"min_fare\", \"max_fare\", and \"mean_fare\" are the minimum, maximum, and mean of the \"Fare\" column after Min-Max scaling, respectively. \"min_fare\" and \"max_fare\" are rounded to two decimal places, while \"mean_fare\" is rounded to four decimal places.", "keywords": [ "Comprehensive Data Preprocessing", "Feature Engineering" ], "level": "hard", "original_id": 178 }
No context is necessary for this task.	Calculate the Pearson correlation coefficient between the age and fare variables for passengers who survived and were in first class.	[['correlation_coefficient', '-0.123']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "Use only passengers that survived and were in the first class. Calculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between age and fare.", "format": "@correlation_coefficient[c_value] where \"c_value\" is a number between -1 and 1, rounded to three decimal places.", "keywords": [ "Summary Statistics", "Correlation Analysis" ], "level": "medium", "original_id": 179 }
No context is necessary for this task.	Perform outlier detection on the fare variable for each passenger class separately. Use the Z-score method and determine the number of outliers in each class.	[['class2_outliers', '7'], ['class1_outliers', '3'], ['class3_outliers', '14']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "Validate outliers using the Z-score method with a threshold of 3. Use separate calculations for each passenger class (1, 2, and 3).", "format": "@class1_outliers[o1_value], @class2_outliers[o2_value], @class3_outliers[o3_value] where \"o1_value\", \"o2_value\", and \"o3_value\" are non-negative integers representing the count of outliers in each class.", "keywords": [ "Outlier Detection" ], "level": "medium", "original_id": 180 }
No context is necessary for this task.	1. Which column(s) contain missing values in the dataset?	[['missing_columns_in_object_type', 'author, urlToImage']]	[ "/data/daeval/fb_articles_20180822_20180829_df.csv" ]	{ "constraints": "This question requires a straightforward identification of columns with missing values in the dataset. Only count the missing values in columns where the data type is 'object' (i.e., strings). Do not include columns of other data types and consider a \"missing value\" as one that is recorded as 'NaN', 'na', 'null', or an empty string in the dataset.", "format": "@missing_columns_in_object_type[missing_column1, missing_column2,…] whereby 'missing_column1', 'missing_column2', etc. are string names of the columns with missing values. The answer should not contain any duplicates and should be sorted alphabetically for easy checking.", "keywords": [ "Comprehensive Data Preprocessing" ], "level": "easy", "original_id": 207 }
No context is necessary for this task.	2. Calculate the mean and standard deviation of the "compound" sentiment score column.	[['compound_mean', '0.141'], ['compound_std', '0.899']]	[ "/data/daeval/fb_articles_20180822_20180829_df.csv" ]	{ "constraints": "Calculate the mean and standard deviation of the 'compound' sentiment score using standard statistical methods. Please use a standard approach and do not use any approximations or assumptions. Note that the 'compound' column contains no missing values according to the scenario information.", "format": "@compound_mean[mean_value]\\n@compound_std[std_value] where 'mean_value' and 'std_value' should be rounded to three decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 208 }
No context is necessary for this task.	3. Is there any correlation between the "neg" and "pos" sentiment score columns? If so, what is the correlation coefficient?	[['correlation_coefficient', '-0.24']]	[ "/data/daeval/fb_articles_20180822_20180829_df.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between 'neg' and 'pos' sentiment scores. If the Pearson correlation coefficient (absolute value) is close to 1, it means that there exists a strong correlation. If it is close to 0, it means that there exists a weak or no correlation. If the coefficient is positive, the correlation is positive; if negative, the correlation is negative.", "format": "@correlation_coefficient[r_value] where 'r_value' is a number between -1 and 1, rounded to two decimal places.", "keywords": [ "Correlation Analysis" ], "level": "medium", "original_id": 209 }
No context is necessary for this task.	1. Identify and remove any outliers in the "neg" sentiment score column using the Z-score method, where Z is defined as (value - mean) / standard deviation. Assume a data point to be an outlier if its Z-score is greater than 3 or less than -3. After removing outliers, calculate the new mean and standard deviation for the "neg" sentiment score column.	[['mean_neg', '0.07'], ['std_dev_neg', '0.04']]	[ "/data/daeval/fb_articles_20180822_20180829_df.csv" ]	{ "constraints": "Z-score is calculated with its general mathematical formula (value - mean) / standard deviation. Consider a data point as an outlier if its Z-score is greater than 3 or less than -3. Do this for the \"neg\" sentiment score column only.", "format": "@mean_neg[mean]\\n@std_dev_neg[std_dev] where \"mean\" and \"std_dev\" are floating-point numbers rounded to two decimal places. Additionally, \"mean\" and \"std_dev\" should be greater than 0 and less than 1 as they mimic sentiment scores.", "keywords": [ "Outlier Detection", "Summary Statistics" ], "level": "hard", "original_id": 210 }
No context is necessary for this task.	Calculate the mean and standard deviation of the abs_diffsel column.	[['std_dev', '1.50'], ['mean', '4.61']]	[ "/data/daeval/ferret-Pitt-2-preinf-lib2-100_sitediffsel.csv" ]	{ "constraints": "{\nThe mean and standard deviation should be calculated directly from the 'abs_diffsel' column.\nDo not remove any outliers or modify the data prior to calculation.\nThe mean and standard deviation should be computed directly from all available data points.\n}", "format": "{\n@mean[mean_value]\n@std_dev[std_dev_value]\nwhere \"mean_value\" is a positive float number, rounded to two decimal places.\nwhere \"std_dev_value\" is a positive float number, rounded to two decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 216 }
No context is necessary for this task.	Find the site identifier(s) with the highest positive_diffsel value.	[['site_identifier', '57']]	[ "/data/daeval/ferret-Pitt-2-preinf-lib2-100_sitediffsel.csv" ]	{ "constraints": "{\nExtract the site identifier corresponding to the highest positive_diffsel value.\nIn the case multiple sites have the same highest positive_diffsel value, list all site identifiers.\nAssume the 'positive_diffsel' column contains only unique values unless specified otherwise.}", "format": "{\n@site_identifier[site_identifier]\nwhere \"site_identifier\" is a string from the 'site' column. If multiple site identifiers should be listed, return them as a single string separated by commas.", "keywords": [ "Distribution Analysis" ], "level": "easy", "original_id": 217 }
No context is necessary for this task.	Calculate the correlation coefficient between the positive_diffsel and negative_diffsel columns.	[['correlation_coefficient', '0.08']]	[ "/data/daeval/ferret-Pitt-2-preinf-lib2-100_sitediffsel.csv" ]	{ "constraints": "{\nCalculate the Pearson correlation coefficient (r) to assess the strength of the linear relationship between positive_diffsel and negative_diffsel. \nDo not remove any outliers or modify the data prior to computation. \nUse all available data points for the computation of the correlation coefficient.}", "format": "{\n@correlation_coefficient[r_value]\nwhere \"r_value\" is a float number between -1 and 1, rounded to two decimal places.", "keywords": [ "Correlation Analysis" ], "level": "easy", "original_id": 218 }
No context is necessary for this task.	Identify the site(s) with outliers in the abs_diffsel column using the interquartile range (IQR) method. An outlier is defined as a value that is below Q1 - 1.5IQR or above Q3 + 1.5IQR. Provide the site identifier(s) and the corresponding absolute difference in selection values for the outliers.	[['outlier_values', '9.03,9.00'], ['site_identifiers', '(HA2)121,326']]	[ "/data/daeval/ferret-Pitt-2-preinf-lib2-100_sitediffsel.csv" ]	{ "constraints": "Use the interquartile range (IQR) method for outlier detection. Calculate the IQR as the difference between the first quartile (Q1) and the third quartile (Q3) of the abs_diffsel column. Consider a value as an outlier if it is below Q1 - 1.5IQR or above Q3 + 1.5IQR.", "format": "@site_identifiers[site_id1,site_id2,...] @outlier_values[value1,value2,...], where \"site_id1\", \"site_id2\", ... are strings representing the identifiers of the sites with detected outliers and \"value1\", \"value2\", ... are the corresponding absolute difference in selection values for the outliers, each rounded to two decimal places.", "keywords": [ "Outlier Detection" ], "level": "medium", "original_id": 219 }
No context is necessary for this task.	Utilize machine learning techniques to classify the sites into two categories based on their positive_diffsel values, with values less than or equal to the mean defined as 'low' selection, and the rest as 'high'. Split the dataset into training and testing sets with an 80:20 ratio using a specified random state of 42. Train a logistic regression model on the training set, and evaluate its performance on the testing set using accuracy as a metric.	[['accuracy_score', '0.98']]	[ "/data/daeval/ferret-Pitt-2-preinf-lib2-100_sitediffsel.csv" ]	{ "constraints": "Use scikit-learn's Logistic Regression for your classifier model, 'liblinear' solver for the Logistic Regression, and a random state of 42 when splitting the data and building the model. All numeric values should be rounded to two decimal places.", "format": "@accuracy_score[float between 0 and 1, inclusive, rounded to two decimal places].", "keywords": [ "Machine Learning", "Distribution Analysis" ], "level": "hard", "original_id": 224 }
No context is necessary for this task.	What is the average duration of a budget year for all departments?	[['average_duration', '364']]	[ "/data/daeval/city_departments_in_current_budget.csv" ]	{ "constraints": "Calculate the duration of each department’s budget year in days, by subtracting the budget_year_start from budget_year_end. Afterwards, calculate the mean of these durations using a built-in Python function.", "format": "@average_duration[days] where \"days\" is a whole number representing the average number of days of budget years across all departments, rounded to the nearest whole number.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 234 }
No context is necessary for this task.	What is the mean batting average of the players in the dataset?	[['mean_batting_average', '0.258']]	[ "/data/daeval/baseball_data.csv" ]	{ "constraints": "Consider only the players who have a recorded batting average. Do not include the one player with a missing batting average into your calculation. Report your answer rounded off to three decimal places.", "format": "@mean_batting_average[mean_batting_average] where \"mean_batting_average\" is a float between 0 and 1, rounded to three decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 243 }
No context is necessary for this task.	What is the average number of runs scored by players who are eligible for free agency compared to players who are not eligible for free agency?	[['average_runs_by_not_eligible_for_free_agency', '39.63'], ['average_runs_by_eligible_for_free_agency', '57.41']]	[ "/data/daeval/baseball_data.csv" ]	{ "constraints": "Use the 'groupby' function on the 'indicator_of_free_agency_eligibility' column to group the data by whether a player is eligible for free agency or not. Then use the 'mean' function on the 'number_of_runs' column to find the average number of runs scored by these two groups of players.", "format": "@average_runs_by_eligible_for_free_agency[average_runs]\\n@average_runs_by_not_eligible_for_free_agency[average_runs] where \"average_runs\" is a float rounded to two decimal places specifying the average number of runs scored by players who are eligible for free agency and players who are not eligible for free agency, respectively.", "keywords": [ "Summary Statistics" ], "level": "medium", "original_id": 247 }
No context is necessary for this task.	Create a new feature called "batting_average_minus_on_base_percentage" which represents the difference between a player's batting average and their on-base percentage. Calculate the mean and standard deviation of this new feature.	[['mean', '-0.07'], ['std_dev', '0.03']]	[ "/data/daeval/baseball_data.csv" ]	{ "constraints": "To calculate the new feature, subtract each player's on-base percentage from their batting average. Ignore the missing values and areas with null values for batting average or on-base percentage. Calculate both the mean and standard deviation using these new feature values.", "format": "@mean[mean_value] @std_dev[std_dev_value] where \"mean_value\" is the mean of the new feature, and \"std_dev_value\" is the standard deviation of the new feature. Both should be rounded to two decimal places.", "keywords": [ "Feature Engineering", "Summary Statistics" ], "level": "medium", "original_id": 250 }
No context is necessary for this task.	Determine which country's gross domestic product per capita in the year 1992 had the highest skewness among all countries in the dataset.	[['highest_skewness_country', 'Afghanistan']]	[ "/data/daeval/gapminder_gdp_asia.csv" ]	{ "constraints": "Use Python's SciPy library to calculate the skewness of each country's gross domestic product per capita in 1992. Skewness should be calculated with Fisher’s definition, i.e. the one that's adjusted for the normal distribution.", "format": "@highest_skewness_country[country_name] where \"country_name\" is a string", "keywords": [ "Distribution Analysis" ], "level": "medium", "original_id": 252 }
No context is necessary for this task.	Identify any outliers in the gross domestic product per capita data for the year 1982 for all countries. Define an outlier as any data point that falls more than 1.5 times the interquartile range (IQR) below the first quartile or above the third quartile. Report the country or countries which their gdpPercap_1982 values are identified as outliers.	[['outlier_countries', 'Kuwait, Saudi Arabia']]	[ "/data/daeval/gapminder_gdp_asia.csv" ]	{ "constraints": "Use the interquartile range (IQR) rule for outlier detection: a data point is considered an outlier if it falls more than 1.5*IQR below the first quartile (Q1) or above the third quartile (Q3). Don't use any other outlier detection methods or parameters.", "format": "@outlier_countries[list_of_strings] where each string in list_of_strings is the name of a country that is an outlier according to the IQR rule.", "keywords": [ "Outlier Detection" ], "level": "medium", "original_id": 254 }
No context is necessary for this task.	Calculate the mean and standard deviation of the gross domestic product per capita in the year 2007 for all countries in the dataset. Round your answers to 2 decimal places.	[['standard_deviation_gdp2007', '14154.94'], ['mean_gdp2007', '12473.03']]	[ "/data/daeval/gapminder_gdp_asia.csv" ]	{ "constraints": "Perform arithmetic mean and standard deviation calculations on the 'gdpPercap_2007' column of the dataset. Round your answer to two decimal places. Do not use modes, medians, or any other form of average.", "format": "@mean_gdp2007[float], @standard_deviation_gdp2007[float] where each float is a positive number rounded to two decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 255 }
No context is necessary for this task.	Is there any correlation between the TOTUSJH and TOTUSJZ columns in the dataset?	[['correlation_type', 'Positive Correlation'], ['correlation_coefficient', '0.99'], ['p_value', '0.000']]	[ "/data/daeval/3901.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient between the TOTUSJH and TOTUSJZ columns. Use a two-tailed test with a significance level of 0.05 to determine the statistical significance. If the p-value is less than 0.05, report the relationship as either \"Positive Correlation\", \"Negative Correlation\" or \"No Correlation\", based on the sign and magnitude of the correlation coefficient. If the p-value is greater than or equal to 0.05, report \"No Significant Correlation\".", "format": "@correlation_type[A string that should be either \"Positive Correlation\", \"Negative Correlation\", \"No Correlation\", or \"No Significant Correlation\"] @correlation_coefficient[The Pearson correlation coefficient, rounded to two decimal places] @p_value[The p-value of the correlation test, rounded to three decimal places]", "keywords": [ "Correlation Analysis" ], "level": "medium", "original_id": 269 }
No context is necessary for this task.	Create a new feature named "TOTUSJZ_TOTUSJH_RATIO" by dividing the TOTUSJZ column by the TOTUSJH column. Calculate the mean and standard deviation of this new feature.	[['mean_ratio', '22756785531.29'], ['stddev_ratio', '969133356.79']]	[ "/data/daeval/3901.csv" ]	{ "constraints": "Add a small constant (1e-10) to the denominator (TOTUSJH column) to avoid dividing by zero.", "format": "Return 2 values: @mean_ratio[Mean of the TOTUSJZ_TOTUSJH_RATIO column, rounded to two decimal places], @stddev_ratio[Standard deviation of the TOTUSJZ_TOTUSJH_RATIO column, rounded to two decimal places].", "keywords": [ "Feature Engineering", "Summary Statistics" ], "level": "medium", "original_id": 272 }
No context is necessary for this task.	Perform a correlation analysis between the MEANGAM and MEANGBT columns. Additionally, for the correlated variables, identify any outliers in the MEANGAM column using the Z-score method and a threshold of 3 for the absolute Z-score.	[['correlation_coefficient', '0.922'], ['outlier_count', '0'], ['outlier_list', '']]	[ "/data/daeval/3901.csv" ]	{ "constraints": "1. Use the Pearson correlation coefficient to assess the correlation between MEANGAM and MEANGBT columns.\n2. Define outliers as those data points in the MEANGAM column where the absolute Z-score exceeds 3.", "format": "1. @correlation_coefficient[correlation_value] where \"correlation_value\" should be a number between -1 and 1, rounded to 3 decimal places.\n2. @outlier_count[outlier_total] where \"outlier_total\" denotes the total number of identified outliers in the MEANGAM column.\n3. @outlier_list[outlier_values_list] where \"outlier_values_list\" is a list of the identified outlier values in MEANGAM column, rounded to 2 decimal places.", "keywords": [ "Correlation Analysis", "Outlier Detection" ], "level": "hard", "original_id": 273 }
No context is necessary for this task.	Is there any correlation between the MedInd and LarInd columns in the given dataset? If yes, what is the correlation coefficient?	[['correlation_coefficient', '0.7366']]	[ "/data/daeval/veracruz 2016.csv" ]	{ "constraints": "Calculate the Pearson's correlation coefficient (r), a statistical measure that calculates the strength of the relationship between the relative movements of two variables, between the MedInd and LarInd columns. The Pearson's correlation coefficient should be rounded to 4 decimal places.", "format": "@correlation_coefficient[correlation_value] where \"correlation_value\" is a signed numeric value between -1 and 1, rounded to 4 decimal places.", "keywords": [ "Correlation Analysis" ], "level": "medium", "original_id": 277 }
No context is necessary for this task.	Are there any outliers in the Agri column of the dataset? If yes, how would you detect them using Z-scores?	[['outliers_count', '0']]	[ "/data/daeval/veracruz 2016.csv" ]	{ "constraints": "Calculate the Z-scores for the Agri column values. Any data point that has a Z-score greater than 3 or less than -3 should be considered as an outlier.", "format": "@outliers_count[outliers_value] where \"outliers_value\" is a non-negative integer representing the count of outliers detected based on the Z-score calculation.", "keywords": [ "Outlier Detection" ], "level": "easy", "original_id": 278 }
No context is necessary for this task.	Perform correlation analysis on the given dataset to determine if there is any relationship between the Agri and Residential columns. Additionally, explore the distribution of the Agri column and identify any outliers using z-score as the outlier detection method. Treat any value which has z-score above 3 as an outlier.	[['number_of_outliers', '0'], ['correlation_coefficient', '-0.17']]	[ "/data/daeval/veracruz 2016.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient to assess the linear relationship between Agri and Residential columns. Treat a value as an outlier if the z-score is above 3. Do not consider any other method for outlier detection. Use scipy's pearsonr method for correlation calculation.", "format": "@correlation_coefficient[r_value]\\n@number_of_outliers[number_of_outliers]\\n where \"r_value\" is a number between -1 and 1, rounded to two decimal places. \"number_of_outliers\" is an integer representing the number of outliers found in the Agri column.", "keywords": [ "Correlation Analysis", "Distribution Analysis", "Outlier Detection" ], "level": "hard", "original_id": 282 }
No context is necessary for this task.	Perform distribution analysis on the age and fare variables separately, then calculate and compare the skewness and kurtosis values for each. Additionally, count the number of values within one standard deviation from the mean, for both age and fare.	[['fare_kurtosis', '33.20'], ['age_values_within_one_stdev', '516'], ['fare_skewness', '4.78'], ['fare_values_within_one_stdev', '818'], ['age_skewness', '0.39'], ['age_kurtosis', '0.17']]	[ "/data/daeval/titanic.csv" ]	{ "constraints": "Use Python's scipy library for your analyses. Calculate skewness and kurtosis values using the scipy.stats.skew() and scipy.stats.kurtosis() functions, respectively, with the default settings. Count the number of values within one standard deviation from the mean by applying standard formula: mean - stdev <= x <= mean + stdev.", "format": "@age_skewness[skewness_value] \n@age_kurtosis[kurtosis_value] \n@age_values_within_one_stdev[number]\n@fare_skewness[skewness_value] \n@fare_kurtosis[kurtosis_value] \n@fare_values_within_one_stdev[number]\nwhere \"skewness_value\", \"kurtosis_value\" are floats with two decimals, \"number\" is a positive integer.", "keywords": [ "Distribution Analysis", "Summary Statistics" ], "level": "hard", "original_id": 309 }
No context is necessary for this task.	What is the mean of the EVENTMSGTYPE column?	[['mean_eventmsgtype', '3.98']]	[ "/data/daeval/0020200722.csv" ]	{ "constraints": "To calculate the arithmetic mean of all the observations in the EVENTMSGTYPE column. Ignore any missing values or outliers when calculating the mean.", "format": "@mean_eventmsgtype[mean], rounded to two decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 320 }
No context is necessary for this task.	Are there any missing values in the dataset? If so, which column has the highest number of missing values?	[['max_missing_values', 'NEUTRALDESCRIPTION']]	[ "/data/daeval/0020200722.csv" ]	{ "constraints": "Columns with missing values: HOMEDESCRIPTION, NEUTRALDESCRIPTION, VISITORDESCRIPTION, SCORE, SCOREMARGIN.", "format": "@max_missing_values[column_name] The output should be a string containing the name of the column with the highest number of missing values.", "keywords": [ "Comprehensive Data Preprocessing" ], "level": "easy", "original_id": 324 }
No context is necessary for this task.	Create a new feature named "event_hour" that represents the hour of the day (in 24-hour format) when each event occurred. Perform a correlation analysis to determine if there is a relationship between the event hour and the event type (EVENTMSGTYPE).	[['relationship', 'none'], ['correlation_coefficient', '0.08'], ['p_value', '0.0749']]	[ "/data/daeval/0020200722.csv" ]	{ "constraints": "{\nExtract the hour from the WCTIMESTRING column using string manipulation functions.\nCalculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between event_hour and EVENTMSGTYPE.\nAssess the significance of the correlation using a two-tailed test with a significance level (alpha) of 0.05.\nReport the p-value associated with the correlation test.\nConsider the relationship to be positive if the p-value is less than 0.05 and the correlation coefficient is greater than or equal to 0.5.\nConsider the relationship to be negative if the p-value is less than 0.05 and the correlation coefficient is less than or equal to -0.5.\nIf the p-value is greater than or equal to 0.05, report that there is no significant correlation.\n}", "format": "{\n@correlation_coefficient[r_value]\n@p_value[p_value]\n@relationship[relationship]\nwhere \"r_value\" is a number between -1 and 1, rounded to two decimal places.\nwhere \"p_value\" is a number between 0 and 1, rounded to four decimal places.\nwhere \"relationship\" is a string that can either be \"positive\", \"negative\", or \"none\" based on the conditions specified in the constraints.", "keywords": [ "Feature Engineering", "Correlation Analysis" ], "level": "hard", "original_id": 326 }
No context is necessary for this task.	2. Is the distribution of the median sold price per square foot skewed? If yes, is it positively or negatively skewed?	[['skewness_type', 'Positive Skewness'], ['skewness_coefficient', '0.08']]	[ "/data/daeval/Zip_MedianSoldPricePerSqft_AllHomes.csv" ]	{ "constraints": "For determining the skewness, consider only non-null values. Use the Fisher-Pearson standardized moment coefficient for assessing the skewness. A skewness value > 0 means that there is more weight in the right tail of the distribution (positive skewness). A skewness value < 0 means that there is more weight in the left tail of the distribution (negative skewness). Calculate the skewness up to two decimal places.", "format": "@skewness_coefficient[skewness_coefficient]\n@skewness_type[skewness_type]\nwhere \"skewness_coefficient\" is a number greater than or equal to -1 and less than or equal to 1, rounded to two decimal places.\nwhere \"skewness_type\" is a string that could be either \"Positive Skewness\", \"Negative Skewness\", or \"No Skewness\".", "keywords": [ "Distribution Analysis" ], "level": "medium", "original_id": 337 }
No context is necessary for this task.	3. Is there a correlation between the size rank of a region and the median sold price per square foot? If yes, is it a positive or negative correlation?	[['correlation_type', 'Positive Correlation'], ['correlation_coefficient', '0.178']]	[ "/data/daeval/Zip_MedianSoldPricePerSqft_AllHomes.csv" ]	{ "constraints": "Calculate the Pearson correlation between the variables size rank and median sold price per square foot, considering only non-null values. A correlation value > 0 indicates a positive correlation, while a value < 0 indicates a negative correlation. A correlation value close to zero indicates no correlation. Calculate the correlation coefficient up to three decimal places.", "format": "@correlation_coefficient[correlation_coefficient]\n@correlation_type[correlation_type]\nwhere \"correlation_coefficient\" is a number between -1 and 1, rounded to three decimal places.\nwhere \"correlation_type\" is a string that could be either \"Positive Correlation\", \"Negative Correlation\", or \"No Correlation\".", "keywords": [ "Correlation Analysis" ], "level": "medium", "original_id": 338 }
No context is necessary for this task.	Calculate the mean age of the passengers.	[['mean_age', '1.10']]	[ "/data/daeval/test_x.csv" ]	{ "constraints": "The mean should be calculated on the full 'Age' column with no filtering. Use the default parameter values for pandas.DataFrame.mean method; in particular, ignore NA/null values and compute the arithmetic mean along the specified axis.", "format": "@mean_age[mean_age] where \"mean_age\" is a floating point number representing the calculated mean age, rounded to two decimal places.", "keywords": [ "Summary Statistics" ], "level": "easy", "original_id": 349 }
No context is necessary for this task.	Check if the Fare column follows a normal distribution.	[['is_normal', 'False']]	[ "/data/daeval/test_x.csv" ]	{ "constraints": "Perform a Shapiro-Wilk test for normality on the 'Fare' column. Use a significance level (alpha) of 0.05 to determine if the 'Fare' column is normally distributed. The 'Fare' column is considered to be normally distributed if the p-value from the Shapiro-Wilk test is greater than 0.05.", "format": "@is_normal[is_normal] where \"is_normal\" is a boolean value: True means the 'Fare' column follows a normal distribution; False means it does not follow a normal distribution.", "keywords": [ "Distribution Analysis" ], "level": "easy", "original_id": 350 }
No context is necessary for this task.	Determine the correlation coefficient between Age and Fare.	[['correlation_coefficient', '0.32']]	[ "/data/daeval/test_x.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient between 'Age' and 'Fare'. Use pandas.DataFrame.corr method with the 'pearson' method. Ignore NA/null values.", "format": "@correlation_coefficient[correlation_coefficient] where \"correlation_coefficient\" is a floating point number representing the calculated correlation coefficient, rounded to two decimal places.", "keywords": [ "Correlation Analysis" ], "level": "easy", "original_id": 351 }
No context is necessary for this task.	Create a new feature "FamilySize" by summing the IsAlone column with the number of siblings/spouses and number of parents/children on board.	[['average_familysize', '2.6']]	[ "/data/daeval/test_x.csv" ]	{ "constraints": "Assume each passenger has at least one sibling/spouse and one parent/child on board, therefore, FamilySize = IsAlone + 1 (for sibling or spouse) + 1 (for parent or child).\nCompute the average FamilySize and round to one decimal place.", "format": "@average_familysize[avg_family_size]\nwhere \"avg_family_size\" is a number rounded to one decimal place.", "keywords": [ "Feature Engineering" ], "level": "easy", "original_id": 354 }
No context is necessary for this task.	Perform a linear regression analysis to predict fare based on age and passenger class.	[['relationship_age', 'not significant'], ['relationship_pclass', 'significant'], ['coef_pclass', '-0.98'], ['coef_age', '-0.05']]	[ "/data/daeval/test_x.csv" ]	{ "constraints": "{\nUse the simple linear regression model where Fare is the dependent variable and Age and Pclass are the independent variables.\nConsider the relationship to be significant if the p-value is less than 0.05 for both variables (Age and Pclass).\nIf the p-value is greater than or equal to 0.05 for either variable, report that there is no significant relationship.\n}", "format": "{\n@coef_age[coef_age]\n@coef_pclass[coef_pclass]\n@relationship_age[relationship_age]\n@relationship_pclass[relationship_pclass]\nwhere \"coef_age\" and \"coef_pclass\" are the regression coefficients for the variables Age and Pclass respectively, rounded to two decimal places.\nwhere \"relationship_age\" and \"relationship_pclass\" are strings that can either be \"significant\" or \"not significant\" based on the conditions specified in the constraints.", "keywords": [ "Correlation Analysis", "Machine Learning" ], "level": "hard", "original_id": 355 }
No context is necessary for this task.	Check if the distribution of wind speed in the weather dataset is skewed.	[['skewness_value', '0.83'], ['skewness_type', 'positive']]	[ "/data/daeval/weather_train.csv" ]	{ "constraints": "For missing values in the \"wind speed\" column, use the 'dropna' method to remove these data points before calculations.\nDetermine the skewness using Pearson's First Coefficient of Skewness. \nReport whether the distribution is positively skewed, negatively skewed, or symmetric based on the obtained skewness value. \nAssume the distribution to be positively skewed if skewness value is > 0, negatively skewed if skewness is < 0, and symmetric if skewness is 0.", "format": "@skewness_value[skew_value]\n@skewness_type[type_value]\nwhere \"skew_value\" is a float number rounded to 2 decimal places.\nwhere \"type_value\" is a string that can be either \"positive\", \"negative\", or \"symmetric\" based on the conditions specified in the constraints.", "keywords": [ "Distribution Analysis" ], "level": "easy", "original_id": 359 }
No context is necessary for this task.	Determine the correlation coefficient between temperature and humidity in the weather dataset.	[['correlation_strength', 'moderate'], ['correlation_coefficient', '-0.64']]	[ "/data/daeval/weather_train.csv" ]	{ "constraints": "For missing values in either the \"temperature\" or \"humidity\" columns, use the 'dropna' method to remove these datapoints before calculations.\nCalculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between temperature and humidity.\nConsider a correlation to be strong if the absolute value of r is greater than or equal to 0.7, moderate if it is between 0.3 and 0.7, and weak if it is less than 0.3.", "format": "@correlation_coefficient[r_value]\n@correlation_strength[strength_value]\nwhere \"r_value\" is a float number between -1 and 1, rounded to 2 decimal places.\nwhere \"strength_value\" is a string that can be either \"strong\", \"moderate\", or \"weak\" based on the conditions specified in the constraints.", "keywords": [ "Correlation Analysis" ], "level": "medium", "original_id": 360 }
No context is necessary for this task.	Identify and remove outliers in the wind speed column of the weather dataset. Use the Z-score method to detect outliers with a threshold of 3 and create a new dataframe without the outlier values.	[['outlier_count', '0']]	[ "/data/daeval/weather_train.csv" ]	{ "constraints": "{\nUse a Z-score threshold of 3 for outlier identification.\nIf the Z-score of a value is higher than 3 or lower than -3, consider it as an outlier.\nAfter outlier detection, drop these rows and create a new dataframe.\n}", "format": "{\n@outlier_count[integer]\nwhere \"integer\" represents the total count of outliers detected.", "keywords": [ "Outlier Detection", "Comprehensive Data Preprocessing" ], "level": "medium", "original_id": 361 }
No context is necessary for this task.	Train a machine learning model to predict the amount of sunlight (sun column) based on the temperature, humidity, and wind speed columns. Use a simple linear regression model. Split the dataset into a 70-30 training-testing split, and evaluate the model's performance using the mean squared error.	[['mean_squared_error', '1.18']]	[ "/data/daeval/weather_train.csv" ]	{ "constraints": "{\nUse a simple linear regression model for training.\nSplit the data into training and testing sets in a 70-30 ratio.\nEvaluate the model using mean squared error (make sure your mean squared error is not negative).\n}", "format": "{\n@mean_squared_error[mse]\nwhere \"mse\" is a number rounded to 2 decimal places representing the mean squared error of the predictions.", "keywords": [ "Machine Learning" ], "level": "hard", "original_id": 363 }
No context is necessary for this task.	2. Perform a distribution analysis on the "Trips over the past 24-hours (midnight to 11:59pm)" column. Determine if the distribution adheres to a normal distribution or it exhibits skewness, heavy tails, or bimodality.	[['skewness', '0.1520'], ['kurtosis', '-1.1336'], ['shapiro_w', '0.9543'], ['p_value', '0.0027']]	[ "/data/daeval/2014_q4.csv" ]	{ "constraints": "Use Python's scipy library to perform a Shapiro-Wilk test to check for normality. The Shapiro-Wilk test tests the null hypothesis that the data was drawn from a normal distribution. For skewness and kurtosis use Python's scipy library. Results for skewness and kurtosis are defined as 'heavy' if they fall outside the range of -0.5 to 0.5.", "format": "@shapiro_w[test_statistic], @p_value[p_value], @skewness[skewness_value], @kurtosis[kurtosis_value] where each answer is a floating number to four decimal places. If p-value is less than 0.05, the distribution is not normal.", "keywords": [ "Distribution Analysis" ], "level": "medium", "original_id": 375 }
No context is necessary for this task.	Is there a correlation between the fare paid by the passenger and their age? If so, is it a linear or nonlinear correlation?	[['correlation_coefficient', '0.10'], ['relationship_type', 'nonlinear'], ['p_value', '0.0102']]	[ "/data/daeval/titanic_train.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between 'Fare' and 'Age'.\nAssess the significance of the correlation using a two-tailed test with a significance level (alpha) of 0.05.\nReport the p-value associated with the correlation test.\nConsider the relationship to be linear if the p-value is less than 0.05 and the absolute value of r is greater than or equal to 0.5.\nConsider the relationship to be nonlinear if the p-value is less than 0.05 and the absolute value of r is less than 0.5.\nIf the p-value is greater than or equal to 0.05, report that there is no significant correlation.\nIgnore the null values in 'Age' while calculating the correlation.", "format": "@correlation_coefficient[r_value]\n@p_value[p_value]\n@relationship_type[relationship_type] \nwhere \"r_value\" is a number between -1 and 1, rounded to two decimal places.\nwhere \"p_value\" is a number between 0 and 1, rounded to four decimal places.\nwhere \"relationship_type\" is a string that can either be \"linear\", \"nonlinear\", or \"none\" based on the conditions specified in the constraints.", "keywords": [ "Correlation Analysis" ], "level": "medium", "original_id": 408 }
No context is necessary for this task.	How many missing values are there in the "Cabin" column?	[['missing_values', '687']]	[ "/data/daeval/titanic_train.csv" ]	{ "constraints": "Count the number of missing values in the 'Cabin' column in the dataset. Treat null values as missing values.", "format": "@missing_values[missing_values] where \"missing_values\" is an integer.", "keywords": [ "Comprehensive Data Preprocessing" ], "level": "easy", "original_id": 409 }
No context is necessary for this task.	What is the distribution of ages among the male passengers who did not survive? Is it significantly different from the distribution of ages among the female passengers who did not survive?	[['is_significantly_different', 'True']]	[ "/data/daeval/titanic_train.csv" ]	{ "constraints": "Calculating the distribution of ages should use a Kernel Density Estimation (KDE) method. Perform a two-sample Kolmogorov-Smirnov test to compare the distributions. Use a significance level (alpha) of 0.05. If the p-value is less than 0.05, conclude the distributions are significantly different. If the p-value is greater than or equal to 0.05, conclude the distributions are not significantly different.", "format": "@is_significantly_different[answer] where \"answer\" is a boolean indicating the result of the test. For example, if the distributions are significantly different, the answer should be \"True\". If not, the answer should be \"False\".", "keywords": [ "Distribution Analysis" ], "level": "medium", "original_id": 410 }
No context is necessary for this task.	Create a new feature called "FamilySize" by adding the "SibSp" and "Parch" columns together. What is the mean "FamilySize" for passengers who survived versus passengers who did not survive?	[['mean_familysize_survived', '0.94'], ['mean_familysize_did_not_survive', '0.88']]	[ "/data/daeval/titanic_train.csv" ]	{ "constraints": "Calculate the mean of \"FamilySize\" separately for the passengers who survived and the passengers who did not survive. \"FamilySize\" should be an integer value. The mean should be calculated rounding up to two decimal places.", "format": "@mean_familysize_survived[answer1] @mean_familysize_did_not_survive[answer2] where \"answer1\" is the mean \"FamilySize\" for passengers who survived and \"answer2\" is the mean \"FamilySize\" for passengers who did not survive. Both results should be rounded to 2 decimal places.", "keywords": [ "Feature Engineering" ], "level": "medium", "original_id": 412 }
No context is necessary for this task.	Is there a correlation between the ticket class (Pclass) and the fare paid by the passengers that embarked from Cherbourg (Embarked = 'C')?	[['correlation_coefficient', '-0.53'], ['relationship_significance', 'significant'], ['p_value', '0.0000']]	[ "/data/daeval/titanic_train.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between Pclass and Fare for passengers who embarked from Cherbourg.\nAssess the significance of the correlation using a two-tailed test with a significance level (alpha) of 0.01.\nReport the p-value associated with the correlation test.\nConsider the relationship to be significant if the p-value is less than 0.01.\nIf the p-value is greater than or equal to 0.01, report that there is no significant correlation.", "format": "@correlation_coefficient[r_value]\n@p_value[p_value]\n@relationship_significance[significance]\nwhere \"r_value\" is a number between -1 and 1, rounded to two decimal places.\nwhere \"p_value\" is a number between 0 and 1, rounded to four decimal places.\nwhere \"significance\" is a string that can either be \"significant\" or \"not significant\" based on the conditions specified in the constraints.", "keywords": [ "Correlation Analysis", "Comprehensive Data Preprocessing" ], "level": "hard", "original_id": 413 }
No context is necessary for this task.	What is the average age of passengers in each ticket class (Pclass)?	[['first_class_average_age', '38.23'], ['second_class_average_age', '29.88'], ['third_class_average_age', '25.14']]	[ "/data/daeval/titanic_train.csv" ]	{ "constraints": "Calculate the average (mean) age of the passengers in each class separately (Pclass = 1, Pclass = 2, Pclass = 3).\nIgnore the rows with missing age.\nRound the average age to two decimal places.", "format": "@first_class_average_age[average_age_1]\n@second_class_average_age[average_age_2]\n@third_class_average_age[average_age_3]\nwhere \"average_age_1\" is the average age of the first-class passengers, rounded to two decimal places.\nwhere \"average_age_2\" is the average age of the second-class passengers, rounded to two decimal places.\nwhere \"average_age_3\" is the average age of the third-class passengers, rounded to two decimal places.", "keywords": [ "Summary Statistics", "Comprehensive Data Preprocessing" ], "level": "medium", "original_id": 414 }
No context is necessary for this task.	What is the distribution of fare paid by male passengers who survived? Are there any significant differences in the fare paid by male passengers who survived compared to male passengers who did not survive?	[['survived_fare_mean', '40.82'], ['not_survived_fare_std', '32.41'], ['fare_difference_significance', 'significant'], ['not_survived_fare_mean', '21.96'], ['survived_fare_std', '71.36']]	[ "/data/daeval/titanic_train.csv" ]	{ "constraints": "Calculate the mean and standard deviation of fares paid by male passengers who survived and did not survive separately.\nConduct an independent sample t-test to compare the means of these two groups.\nUse a significance level of 0.05.\nReport whether there is a significant difference in the means based on the p-value of the test.", "format": "@survived_fare_mean[mean_survived]\n@survived_fare_std[std_survived]\n@not_survived_fare_mean[mean_not_survived]\n@not_survived_fare_std[std_not_survived]\n@fare_difference_significance[significance]\nwhere \"mean_survived\" is the mean fare of male passengers who survived, rounded to two decimal places.\nwhere \"std_survived\" is the standard deviation of fare of male passengers who survived, rounded to two decimal places.\nwhere \"mean_not_survived\" is the mean fare of male passengers who did not survive, rounded to two decimal places.\nwhere \"std_not_survived\" is the standard deviation of fare of male passengers who did not survive, rounded to two decimal places.\nwhere \"significance\" is a string that can either be \"significant\" or \"not significant\" based on the conditions specified in the constraints.", "keywords": [ "Distribution Analysis", "Comprehensive Data Preprocessing" ], "level": "hard", "original_id": 415 }
No context is necessary for this task.	1. Is there a significant difference in the mean opening prices between weekdays and weekends? Provide statistical evidence to support your answer.	[['weekday_mean_price', '37.30'], ['weekend_mean_price', '38.49'], ['p_value', '0.8463'], ['significance', 'No']]	[ "/data/daeval/bitconnect_price.csv" ]	{ "constraints": "Calculate the mean opening prices separately for weekdays (Monday to Friday) and weekends (Saturday and Sunday). Conduct a two-sample t-test to check if there is a statistically significant difference between these two means. Use a significance level (alpha) of 0.05. If the p-value is less than 0.05, conclude that there is a significant difference. If the p-value is greater than or equal to 0.05, conclude that there is no significant difference.", "format": "@weekday_mean_price[weekday_mean_open_price] @weekend_mean_price[weekend_mean_open_price] @p_value[p_value] @significance[significant_or_not] where \"weekday_mean_open_price\" and \"weekend_mean_open_price\" are numbers rounded to two decimal places. \"p_value\" is a number between 0 and 1, rounded to four decimal places. \"significant_or_not\" is a string that can either be 'Yes' or 'No' depending on whether the p-value is less than 0.05.", "keywords": [ "Summary Statistics", "Distribution Analysis" ], "level": "hard", "original_id": 419 }
No context is necessary for this task.	2. Perform feature engineering on the given dataset to create a new feature 'Volatility' that is calculated using the formula: Volatility = (High Price - Low Price) / Open Price. What is the Pearson correlation coefficient between Volatility and trading volume? Interpret the result.	[['relationship_type', 'none'], ['correlation_coefficient', '-0.09']]	[ "/data/daeval/bitconnect_price.csv" ]	{ "constraints": "Calculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between the new feature 'Volatility' and trading volume. Consider the relationship to be positive if the r value is greater than or equal to 0.5. Consider the relationship to be negative if the r value is less than or equal to -0.5. Otherwise, report that there is no significant correlation.", "format": "@correlation_coefficient[correlation_coefficient], @relationship_type[type] where \"correlation_coefficient\" is a number between -1 and 1, rounded to two decimal places. \"type\" is a string that can be \"positive\", \"negative\", or \"none\" based on the conditions specified in the constraints.", "keywords": [ "Feature Engineering", "Correlation Analysis" ], "level": "hard", "original_id": 423 }
No context is necessary for this task.	1. How many missing values are there in the "max_sust_wind" column?	[['missing_values_count', '24']]	[ "/data/daeval/cost_data_with_errors.csv" ]	{ "constraints": "Count the number of null values only, do not include non-null values that might be rendered as errors or irrelevant data.", "format": "@missing_values_count[number] where \"number\" is an integer representing the count of the missing values in the \"max_sust_wind\" column.", "keywords": [ "Comprehensive Data Preprocessing" ], "level": "easy", "original_id": 425 }
No context is necessary for this task.	2. What is the maximum sustained wind speed recorded during the storm with the highest maximum storm category?	[['max_wind_speed', '156.42']]	[ "/data/daeval/cost_data_with_errors.csv" ]	{ "constraints": "Firstly, identify the storm with the highest maximum storm category, neglecting any ties. If there are multiple storms with the same highest maximum storm category, choose the one that appears first in the given dataset. Then find the maximum sustained wind speed corresponding to this particular storm.", "format": "@max_wind_speed[number] where \"number\" is a float with two decimal places indicating the highest wind speed recorded for the storm with the highest maximum storm category.", "keywords": [ "Summary Statistics", "Correlation Analysis" ], "level": "medium", "original_id": 426 }
No context is necessary for this task.	3. How many storms have null values in the "min_p" column?	[['null_entries_count', '101']]	[ "/data/daeval/cost_data_with_errors.csv" ]	{ "constraints": "Count the number of entries which have the null value in the \"min_p\" column. Only the null values should be counted, and not any irrelevant or erroneous data that might be present.", "format": "@null_entries_count[number] where \"number\" is an integer indicating the count of null entries in the \"min_p\" column.", "keywords": [ "Comprehensive Data Preprocessing" ], "level": "easy", "original_id": 427 }
No context is necessary for this task.	2. Is there a correlation between the maximum storm category achieved by a storm and the recorded damage in USD? If so, what is the strength and direction of the correlation?	[['correlation_coefficient', '0.19'], ['relationship_type', 'nonlinear'], ['p_value', '0.0000']]	[ "/data/daeval/cost_data_with_errors.csv" ]	{ "constraints": "{\nCalculate the Pearson correlation coefficient (r) to assess the strength and direction of the linear relationship between maximum storm category and damage in USD.\nUse a two-tailed test with a significance level (alpha) of 0.05 to assess the significance of the correlation. \nReport the p-value associated with the correlation test.\nIf the p-value is less than 0.05 and the absolute value of r is greater than or equal to 0.5, infer the relationship to be linear.\nIf the p-value is less than 0.05 and the absolute value of r is less than 0.5, infer the relationship to be nonlinear.\nIf the p-value is greater than or equal to 0.05, report that there is no significant correlation.\n}", "format": "{\n@correlation_coefficient[r_value]\n@p_value[p_value]\n@relationship_type[relationship_type]\nwhere \"r_value\" is a number between -1 and 1, rounded to 2 decimal places.\nwhere \"p_value\" is a number between 0 and 1, rounded to 4 decimal places.\nwhere \"relationship_type\" is a string that can either be \"linear\", \"nonlinear\", or \"none\" based on the conditions specified in the constraints.", "keywords": [ "Correlation Analysis" ], "level": "hard", "original_id": 429 }
No context is necessary for this task.	2. Can we predict the maximum sustained wind speed based on the recorded damage in USD and the minimum recorded pressure? What is the performance of the prediction model?	[['mean_squared_error', '263.1896']]	[ "/data/daeval/cost_data_with_errors.csv" ]	{ "constraints": "{\nUse a linear regression model for prediction.\nUse 80% of the dataset for training and the rest for testing.\nUse Mean Squared Error (MSE) as the evaluation metric to assess the model's performance.\nHandle missing values in the \"max_sust_wind\", \"damage_USD\", and \"min_p\" columns by imputing them with their respective column means.\n}", "format": "{\n@mean_squared_error[mse]\nwhere \"mse\" is the mean squared error of the model, rounded to four decimal places.", "keywords": [ "Machine Learning", "Comprehensive Data Preprocessing" ], "level": "hard", "original_id": 432 }

End of preview.

DSGym

DSGym is a unified benchmark and execution framework for evaluating and training data science agents. It provides standardized, executable tasks that require agents to plan, implement, and validate data analyses through interaction with real data files in isolated environments.

Attribution

If you find our work useful, please cite our paper:

@misc{nie2026dsgym,
      title={DSGym: A Holistic Framework for Evaluating and Training Data Science Agents}, 
      author={Fan Nie and Junlin Wang and Harper Hua and Federico Bianchi and Yongchan Kwon and Zhenting Qi and Owen Queen and Shang Zhu and James Zou},
      year={2026},
      eprint={2601.16344},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2601.16344}, 
}

Downloads last month: 12

Paper for DSGym/DSGym-Tasks

DSGym: A Holistic Framework for Evaluating and Training Data Science Agents

Paper • 2601.16344 • Published 19 days ago • 11