{ "cells": [ { "cell_type": "markdown", "id": "1aaac45b", "metadata": {}, "source": [ "# A2 Report: Deep Squat Movement Assessment\n", "\n", "**Team Members:** Reem, Rasa, Amol (Scrum master for this week), Kalle\n", "\n", "---" ] }, { "cell_type": "markdown", "id": "4e6e0926", "metadata": {}, "source": [ "## 1. Problem Statement\n", "\n", "The goal is to build an ML model that automatically tests Deep Squat movement quality. The input has 40 preprocessed features from 3D skeleton data taken by a Kinect camera. The output is an AimoScore ranging from 0 to 1.0, where higher scores indicate better movement quality." ] }, { "cell_type": "markdown", "id": "7a871603", "metadata": {}, "source": [ "## 2. Dataset Overview\n", "\n", "The dataset contains about 2000 deep squat assessments split into 80% training and 20% test sets." ] }, { "cell_type": "code", "execution_count": 7, "id": "f201d267", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Dataset\n", "Training samples: 1675\n", "Test samples: 419\n", "Total features: 40\n", "\n", "Feature types:\n", " - FMS Angle deviations (body joints): 13 features\n", " - NASM deviations (mvmt patterns): 25 features\n", " - Time deviations (phases of squat): 2 features\n", "\n", "Target: AimoScore (range: 0.01 - 0.99)\n" ] } ], "source": [ "import pandas as pd\n", "import numpy as np\n", "import os\n", "\n", "# Load dataset\n", "REPO_ROOT = os.path.abspath(os.path.join(os.getcwd(), \"..\"))\n", "DATA_DIR = os.path.join(REPO_ROOT, \"Datasets_all\")\n", "\n", "# Load original dataset (with duplicates)\n", "train_df = pd.read_csv(os.path.join(DATA_DIR, \"A2_dataset_80.csv\"))\n", "test_df = pd.read_csv(os.path.join(DATA_DIR, \"A2_dataset_20.csv\"))\n", "\n", "print(\"Dataset\")\n", "print(f\"Training samples: {len(train_df)}\")\n", "print(f\"Test samples: {len(test_df)}\")\n", "print(f\"Total features: {len(train_df.columns) - 2}\") # -2 for AimoScore and EstimatedScore\n", "print(f\"\\nFeature types:\")\n", "print(f\" - FMS Angle deviations (body joints): 13 features\")\n", "print(f\" - NASM deviations (mvmt patterns): 25 features\")\n", "print(f\" - Time deviations (phases of squat): 2 features\")\n", "print(f\"\\nTarget: AimoScore (range: {train_df['AimoScore'].min():.2f} - {train_df['AimoScore'].max():.2f})\")" ] }, { "cell_type": "code", "execution_count": 8, "id": "287985ce", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "\n", "# Plot target distribution\n", "fig, axes = plt.subplots(1, 2, figsize=(12, 4))\n", "\n", "# Target distribution\n", "sns.histplot(train_df['AimoScore'], bins=30, kde=True, ax=axes[0], color='steelblue')\n", "axes[0].set_title('Distribution of AimoScore (Target)')\n", "axes[0].set_xlabel('AimoScore')\n", "\n", "# Box plot\n", "sns.boxplot(x=train_df['AimoScore'], ax=axes[1], color='steelblue')\n", "axes[1].set_title('AimoScore Boxplot')\n", "\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "ba32f22e", "metadata": {}, "source": [ "## 3. Domain Insights\n", "\n", "### 3.1 Duplicate Features\n", "\n", "Some FMS and NASM features measure the same angles and so are identical. We find and remove these duplicates during preprocessing to avoid redundancy and multicollinearity issues." ] }, { "cell_type": "code", "execution_count": 9, "id": "1d45bd6f", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Original features: 40\n", "Duplicate features found: 5\n", "Features after removal: 35\n", "\n", "Duplicate columns removed: ['No_1_NASM_Deviation', 'No_2_NASM_Deviation', 'No_3_NASM_Deviation', 'No_4_NASM_Deviation', 'No_5_NASM_Deviation']\n" ] } ], "source": [ "# Check for duplicate features\n", "X = train_df.drop(columns=['AimoScore', 'EstimatedScore'])\n", "\n", "# Find duplicate columns\n", "duplicates = X.T.duplicated()\n", "duplicate_cols = X.columns[duplicates].tolist()\n", "\n", "print(f\"Original features: 40\")\n", "print(f\"Duplicate features found: {len(duplicate_cols)}\")\n", "print(f\"Features after removal: {40 - len(duplicate_cols)}\")\n", "print(f\"\\nDuplicate columns removed: {duplicate_cols}\")" ] }, { "cell_type": "markdown", "id": "d52959a8", "metadata": {}, "source": [ "### 3.2 Left-Right Symmetry Constraints\n", "\n", "Good squat form should be symmetric on the left and right. We enforce this by averaging left-right feature pairs during training.\n", "\n", "Symmetric FMS pairs: (4,6), (5,7), (8,11), (9,12), (10,13) \n", "Symmetric NASM pairs: (14,15), (17,18), (21,22), (24,25), (26,27), (28,29), (31,32)" ] }, { "cell_type": "markdown", "id": "a33098c9", "metadata": {}, "source": [ "### 3.3 Expert-Defined Feature Weights\n", "\n", "Not all features are equally important for scoring. Domain experts assign weights based on importance." ] }, { "cell_type": "markdown", "id": "8856189e", "metadata": {}, "source": [ "## 4. ML Pipeline Architecture\n", "\n", "We built three sklearn transformers to encode domain knowledge: CorrelationFilter removes very correlated features, SymmetryConstraintTransformer averages left-right pairs, and FeatureWeightMultiplier applies expert weights.\n", "\n", "For outlier removal approach: IQR method to remove samples outside 1.5×IQR, then Cook's Distance to find influential points with threshold 4/n." ] }, { "cell_type": "markdown", "id": "8314bb8e", "metadata": {}, "source": [ "## 5. Model Variants and Evaluation\n", "\n", "We tested 8 model variants using 5-fold cross validation. Each variant combines different preprocessing steps and regularization techniques to find the best performing configuration." ] }, { "cell_type": "code", "execution_count": 10, "id": "9d5b6b9c", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " Variant Technique CV R² (approx)\n", " baseline StandardScaler + LinearRegression ~0.52\n", " corr_filter CorrelationFilter(0.99) + LR ~0.52\n", "feature_selection_lasso LassoCV feature selection + LR ~0.50\n", " combined_poly2_ridge PolynomialFeatures(2) + Ridge ~0.53\n", " symmetry_constraint SymmetryConstraint + LR ~0.54\n", " weighted_features FeatureWeights + LR ~0.55\n", " combined_final CorrFilter + Symmetry + Weights + Ridge ~0.58\n", " outlier_iqr_plus_cooks IQR + Cooks Distance outlier removal + LR ~0.59\n" ] } ], "source": [ "# Model variants comparison\n", "variants_data = {\n", " 'Variant': [\n", " 'baseline',\n", " 'corr_filter',\n", " 'feature_selection_lasso',\n", " 'combined_poly2_ridge',\n", " 'symmetry_constraint',\n", " 'weighted_features',\n", " 'combined_final',\n", " 'outlier_iqr_plus_cooks'\n", " ],\n", " 'Technique': [\n", " 'StandardScaler + LinearRegression',\n", " 'CorrelationFilter(0.99) + LR',\n", " 'LassoCV feature selection + LR',\n", " 'PolynomialFeatures(2) + Ridge',\n", " 'SymmetryConstraint + LR',\n", " 'FeatureWeights + LR',\n", " 'CorrFilter + Symmetry + Weights + Ridge',\n", " 'IQR + Cooks Distance outlier removal + LR'\n", " ],\n", " 'CV R² (approx)': [\n", " '~0.52',\n", " '~0.52',\n", " '~0.50',\n", " '~0.53',\n", " '~0.54',\n", " '~0.55',\n", " '~0.58',\n", " '~0.59'\n", " ]\n", "}\n", "\n", "variants_df = pd.DataFrame(variants_data)\n", "print(variants_df.to_string(index=False))" ] }, { "cell_type": "markdown", "id": "fb2422cb", "metadata": {}, "source": [ "We evaluate each variant using RMSE, MAE, R², and Pearson correlation. The outlier removal variant achieved the best cross-validation performance and was selected as our champion model." ] }, { "cell_type": "markdown", "id": "0a1ec12c", "metadata": {}, "source": [ "## 6. Champion Model Results\n", "\n", "The champion model uses IQR and Cook's Distance outlier removal followed by standard scaling and linear regression. Below we load the saved model and display its performance metrics." ] }, { "cell_type": "code", "execution_count": 11, "id": "287ccafc", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Champion model performance\n", "\n", "Features used: 35\n", "\n", "Test Set Metrics:\n", " RMSE: 0.1493\n", " MAE: 0.1140\n", " R2: 0.5891\n", " CORRELATION: 0.7676\n", "\n", "Cross-Validation Metrics:\n", " cv_rmse_mean: 0.1568\n", " cv_r2_mean: 0.5590\n", " cv_corr_mean: 0.7504\n" ] } ], "source": [ "import pickle\n", "\n", "# Load champion model\n", "model_path = \"models/champion_model_final.pkl\"\n", "with open(model_path, \"rb\") as f:\n", " artifact = pickle.load(f)\n", "\n", "print(\"Champion model performance\")\n", "print(f\"\\nFeatures used: {len(artifact['feature_columns'])}\")\n", "print(f\"\\nTest Set Metrics:\")\n", "for metric, value in artifact['test_metrics'].items():\n", " print(f\" {metric.upper()}: {value:.4f}\")\n", "\n", "print(f\"\\nCross-Validation Metrics:\")\n", "for metric, value in artifact['train_metrics'].items():\n", " print(f\" {metric}: {value:.4f}\")" ] }, { "cell_type": "code", "execution_count": 12, "id": "bacd9ebc", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Visualise predictions on test set\n", "from sklearn.metrics import r2_score\n", "\n", "# Prepare test data - use columns from the trained model\n", "kept_cols = artifact['feature_columns']\n", "# Filter to only columns that exist in test_df\n", "available_cols = [c for c in kept_cols if c in test_df.columns]\n", "X_test = test_df[available_cols].copy()\n", "y_test = test_df['AimoScore']\n", "\n", "# Predict\n", "model = artifact['model']\n", "y_pred = model.predict(X_test)\n", "\n", "# Plot\n", "fig, axes = plt.subplots(1, 2, figsize=(12, 5))\n", "\n", "# Actual vs predicted\n", "axes[0].scatter(y_test, y_pred, alpha=0.5, color='steelblue')\n", "axes[0].plot([y_test.min(), y_test.max()], [y_test.min(), y_test.max()], 'r--', lw=2)\n", "axes[0].set_xlabel('Actual AimoScore')\n", "axes[0].set_ylabel('Predicted AimoScore')\n", "axes[0].set_title(f'Actual vs Predicted (R² = {r2_score(y_test, y_pred):.4f})')\n", "axes[0].grid(True, alpha=0.3)\n", "\n", "# Residuals\n", "residuals = y_test - y_pred\n", "axes[1].hist(residuals, bins=25, edgecolor='black', alpha=0.7, color='steelblue')\n", "axes[1].set_xlabel('Residuals')\n", "axes[1].set_ylabel('Frequency')\n", "axes[1].set_title('Residual Distribution')\n", "axes[1].axvline(x=0, color='red', linestyle='--')\n", "axes[1].grid(True, alpha=0.3)\n", "\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "faef2448", "metadata": {}, "source": [ "## 7. Deployment\n", "\n", "The model is deployed as a Gradio web application on HuggingFace Spaces. Users can input feature values with sliders in three tabs (Angle, NASM, Time deviations) and get a predicted score with interpretation. The app also has a \"Load Random Example\" button to test with real random data from the dataset.\n", "\n", "Deployment URL: https://huggingface.co/spaces/Bachstelze/github_sync" ] }, { "cell_type": "markdown", "id": "1defcf1c", "metadata": {}, "source": [ "## 8. DevOps/MLOps Process\n", "\n", "We use GitHub Actions for continuous deployment. When code is pushed to the main branch, the workflow automatically syncs the repository to HuggingFace Spaces. The file `.github/workflows/push_to_hf_space.yml` handles authentication using the HF_TOKEN secret and force pushes to the HuggingFace space repository." ] }, { "cell_type": "markdown", "id": "befe31ba", "metadata": {}, "source": [ "## 9. Conclusion\n", "\n", "The champion model gets R² of about 0.59, explaining 59% of the variance in expert scores. The remaining variance probably comes from subjective expert judgment and factors not captured by the 40 input features. Another reason could be that different body structures often have different natural forms that the body summumbs to, amongst other domain reasons. " ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.14.0" } }, "nbformat": 4, "nbformat_minor": 5 }