Update dataset card with metadata and EDA plots

README.md

@@ -1,32 +1,22 @@
----
-dataset_info:
-  features:
-  - name: sample_id
-    dtype: string
-  - name: width_cm
-    dtype: float64
-  - name: length_cm
-    dtype: float64
-  - name: blemish_count
-    dtype: int64
-  - name: color
-    dtype: string
-  - name: firmness
-    dtype: int64
-  splits:
-  - name: original
-    num_bytes: 1371
-    num_examples: 30
-  - name: augmented
-    num_bytes: 15572
-    num_examples: 300
-  download_size: 7961
-  dataset_size: 16943
-configs:
-- config_name: default
-  data_files:
-  - split: original
-    path: data/original-*
-  - split: augmented
-    path: data/augmented-*
----
+
+# Dataset Card for Grape Firmness Dataset
+
+## Dataset Metadata
+
+*   **Purpose:** The dataset is intended for training a machine learning model to predict grape firmness based on physical characteristics and blemishes.
+*   **Composition:** The dataset contains measurements of grape width, length, blemish count, color, and firmness.
+*   **Collection Method:** Data was collected manually through physical measurements and visual inspection of grape samples.
+*   **Labels:** The target variable "firmness" is a categorical label with three levels: 1 (Soft), 2 (Medium), and 3 (Firm).
+*   **Splits:**
+    *   `original`: Contains the initial 30 rows of collected data.
+    *   `augmented`: Contains 300 augmented rows generated using SMOTE-NC and numeric jitter to address class imbalance and increase dataset size.
+
+## Exploratory Data Analysis (EDA) - Original Data
+
+Here are some visualizations showing the distribution of firmness and the relationship between physical characteristics and firmness in the original dataset.
+
+![Firmness Distribution in Original Data](images/firmness_distribution_original.png)
+
+![Width vs Firmness in Original Data](images/width_vs_firmness_original.png)
+
+![Length vs Firmness in Original Data](images/length_vs_firmness_original.png)