Skip to content

README.md

Latest commit

 

History

History
90 lines (68 loc) · 3.13 KB

README.md

File metadata and controls

90 lines (68 loc) · 3.13 KB
Raw

⏰ Project 2: Time-Series Model Tracker

🌟 Overview

Dive into Time-Series Model Tracker, where you’ll analyze ML prediction trends over time using DateTime Functions, Window Functions, and Aggregations! This project tracks model performance across months, perfect for your ML-Interview-Preparation-Hub portfolio to show off time-series skills. 🚀

📊 Project Details

  • Scenario: You’re an ML analyst monitoring prediction scores. Your task is to aggregate scores by month, calculate running averages, and identify top models.
  • Goals: Process time-series data, compute trends, and export a report.
  • SQL Skills: DateTime Functions (DATE_TRUNC), Window Functions (RANK, AVG), DQL (SELECT, GROUP BY).

📋 Tasks

  1. Use the predictions table (prediction_id INT, model_id INT, score FLOAT, prediction_date DATE, model_name VARCHAR).
  2. Insert sample data for 2025 predictions.
  3. Aggregate average score by month using DateTime Functions.
  4. Calculate a running average score per model with Window Functions.
  5. Rank models by monthly score.
  6. Export results to a CSV for ML reporting.
  7. Document the process for your portfolio.

💻 Sample Solution

-- Task 1: Table exists (predictions)

-- Task 2: Insert Sample Data
INSERT INTO predictions VALUES
(1, 101, 0.85, '2025-01-15', 'BERT_v1'),
(2, 102, 0.90, '2025-01-20', 'XGBoost_v2'),
(3, 101, 0.87, '2025-02-10', 'BERT_v1'),
(4, 102, 0.88, '2025-02-15', 'XGBoost_v2'),
(5, 103, 0.92, '2025-03-05', 'LSTM_v3'),
(6, 101, 0.86, '2025-03-20', 'BERT_v1'),
(7, 102, 0.91, '2025-04-01', 'XGBoost_v2'),
(8, 103, 0.89, '2025-04-10', 'LSTM_v3');

-- Task 3: Monthly Aggregation
CREATE TABLE monthly_trends AS
SELECT DATE_TRUNC('month', prediction_date) AS month,
       model_id,
       AVG(score) AS avg_score
FROM predictions
GROUP BY month, model_id;

-- Task 4: Running Average
SELECT model_id,
       prediction_date,
       score,
       AVG(score) OVER (PARTITION BY model_id ORDER BY prediction_date) AS running_avg
FROM predictions;

-- Task 5: Rank Models
SELECT month,
       model_id,
       avg_score,
       RANK() OVER (PARTITION BY month ORDER BY avg_score DESC) AS score_rank
FROM monthly_trends;

-- Task 6: Export (Pseudo-code)
-- COPY (SELECT month, model_id, avg_score FROM monthly_trends) TO 'monthly_trends.csv' WITH CSV HEADER;

-- Task 7: Document (See Portfolio Tips)

Explanation:

  • DML inserts time-series data.
  • DateTime Functions (DATE_TRUNC) group by month.
  • Window Functions compute running averages and ranks.
  • DQL creates trend tables for analysis.

🎯 Expected Outcome

  • A monthly_trends table with aggregated scores.
  • A CSV file (monthly_trends.csv) with model performance.
  • A portfolio-ready report showcasing time-series ML skills.

💼 Portfolio Tips

  • Add monthly_trends.csv to irohanportfolio.netlify.app.
  • Write a README explaining trends and ML use (e.g., model monitoring).
  • Plot trends with Python’s matplotlib (e.g., score vs. month).
  • Tag skills: DateTime Functions, Window Functions, DQL.

🔥 Challenge

Add a CTE to compute month-over-month score changes for each model.