ContextLens 🔎

Live Application

Launch ContextLens

Built With

• Python
• Streamlit
• Pandas
• NumPy
• Scikit-learn

Video Demonstration

Click the thumbnail to watch the complete demonstration.

Application Preview

Test

ContextLens is a context-aware machine-learning explorer for tabular CSV data.
It profiles a dataset, explains its modelling context, flags common data risks, trains reproducible baseline pipelines, compares appropriate metrics, and exports a Markdown analysis report.

Built as a portfolio project connecting context-aware computing, data analysis, explainable modelling, and practical Python application development.

What it does

• Upload any tabular CSV or open the included sample dataset.
• Select a target and auto-detect classification versus regression.
• Identify missingness, duplicates, constant columns, possible identifiers, high-cardinality categories, class imbalance, small samples, and high-dimensional settings.
• Compare leakage-safe scikit-learn pipelines.
• Use macro-F1 for classification and RMSE for regression.
• Inspect confusion matrices, per-class reports, and global feature importance.
• Export a portable Markdown report.

Why “context-aware”?

ContextLens does not claim that a heuristic is artificial intelligence. Its context engine uses explicit rules that can be inspected in contextlens/analysis.py. It adapts evaluation guidance and modelling warnings to properties of the selected dataset and target.

Architecture

CSV
 │
 ├── Dataset profiler ──> transparent context rules ──> recommendations
 │
 └── Feature/target split
      └── train-test split
           └── preprocessing inside Pipeline
                ├── numeric: median imputation + scaling
                └── categorical: mode imputation + one-hot encoding
                     └── baseline model comparison
                          └── diagnostics + Markdown report

Preprocessing is fitted only on the training partition through scikit-learn Pipeline and ColumnTransformer, reducing leakage from imputation, scaling, and categorical encoding.

Run locally

git clone https://github.com/anam152/contextlens.git
cd contextlens

python -m venv .venv

Activate the environment:

# Windows PowerShell
.venv\Scripts\Activate.ps1

# macOS/Linux
source .venv/bin/activate

Install and run:

pip install -r requirements.txt
streamlit run app.py

Then open the local URL shown by Streamlit.

Tests

python -m pytest

Deploy on Streamlit Community Cloud

1. Push the repository to GitHub.
2. In Streamlit Community Cloud, create a new app.
3. Select the repository and branch.
4. Set the entry point to app.py.
5. Deploy.

Suggested GitHub topics

streamlit · machine-learning · context-aware-computing · data-science · scikit-learn · explainable-ai · python

Current scope

This is an exploratory baseline application, not a full AutoML platform. Current limitations include:

• one hold-out train/test split rather than repeated cross-validation;
• no time-series-aware splitting;
• global importance rather than local SHAP explanations;
• limited baseline model families;
• no causal, fairness, or domain-validity guarantees.

Roadmap

• Repeated and nested cross-validation
• User-defined group and time-aware splits
• Genetic-algorithm feature selection
• SHAP explanations
• Saved experiment histories
• Domain-context template for healthcare datasets (Pima Indians Diabetes dtatset sample added)
• Domain-context template for IoT datasets
• Downloadable PDF report

Responsible use

Model scores are not proof of real-world validity. High-stakes applications require domain review, representative data, robust validation, fairness and privacy assessment, and appropriate governance.

License

MIT