Python Machine Learning Examples — ARIMA/SARIMAX Time-Series Forecasting and Linear Regression

A collection of beginner-friendly machine learning projects built with Python: a COVID-19 time-series forecast using auto-ARIMA and SARIMAX (pmdarima, statsmodels), and a salary prediction model using scikit-learn linear regression with a separate training and inference workflow.

Projects

1. COVID Forecast

Location: COVID Forecast/Console_Code/CovidForecast.py

A time series forecasting script that predicts the number of daily COVID-19 cases for a country using SARIMAX and auto-ARIMA models.

How It Works

  1. Downloads daily COVID-19 case data via the covid_daily package (if not already cached locally as a CSV)
  2. Performs stationarity testing using the Augmented Dickey-Fuller (ADF) test
  3. Detrends and differences the time series for stationarity
  4. Automatically fits the best ARIMA/SARIMA model using pmdarima.auto_arima
  5. Fits the chosen model using statsmodels.SARIMAX
  6. Generates a static forecast (in-sample) and a future forecast for the next N days
  7. Displays all charts using matplotlib

Configuration

Edit these variables at the top of the script:

Variable Default Description
country 'india' Country to forecast ('india', 'brazil', etc.)
stepCount 50 Number of future days to forecast

Requirements

numpy
pandas
pmdarima
covid_daily
statsmodels
matplotlib

pip install numpy pandas pmdarima covid_daily statsmodels matplotlib

Usage

cd "COVID Forecast/Console_Code"
python CovidForecast.py

The downloaded data is cached as <country>_corvid19.csv in the working directory, so later runs reuse it instead of downloading again.

2. Salary Prediction

Location: Salary/

A linear regression pipeline that predicts salary based on years of experience.

Sub-projects

File Purpose
Model Builder/builder.py Trains the model and saves it to disk
Prediction/prediction.py Loads the saved model and predicts salary from user input

Model Builder Pipeline (builder.py)

  1. Load data from Source/Salary.csv
  2. Remove duplicate rows
  3. Drop columns with more than 50% missing values; fill remaining nulls with column mean
  4. Scale years_experience using StandardScaler
  5. Split into train (80%) and test (20%) sets
  6. Train a LinearRegression model
  7. Evaluate using R² score and Mean Squared Error
  8. Save the model and scaler to Source/salary.pkl and Source/scaler.pkl if R² > 0.95

Prediction Inference (prediction.py)

  1. Load the saved model and scaler from disk
  2. Prompt the user for years of experience
  3. Scale the input using the saved scaler
  4. Predict and display the salary

Requirements

pandas
scikit-learn
joblib
numpy

pip install pandas scikit-learn joblib numpy

Usage

Both scripts use paths relative to the working directory (./Source/...), so run them from the Salary folder. A trained salary.pkl and scaler.pkl are already included in Salary/Source/, so you can run inference without training first.

Train the model:

cd Salary
python "Model Builder/builder.py"

Run inference:

cd Salary
python "Prediction/prediction.py"

Inference session format:

Enter Years of Experience: 5
Predicted Salary for 5.0 years of experience: $<predicted amount>

Folder Structure

ML/
├── COVID Forecast/
│   └── Console_Code/
│       └── CovidForecast.py       # Time series forecasting script
└── Salary/
    ├── Model Builder/
    │   └── builder.py             # Model training pipeline
    ├── Prediction/
    │   └── prediction.py          # Inference script
    └── Source/
        ├── Salary.csv             # Training data
        ├── salary.pkl             # Saved LinearRegression model
        └── scaler.pkl             # Saved StandardScaler

Read all Python Projects and Code Examples guides online · Project overview · All developer projects