Sashank Kurra
← projects

Data / ML · Network Distribution

Supply Chain Forecasting

Time-series forecasting embedded in Tableau

Python · Prophet · TabPy · Tableau · Jupyter · Time-series validation

Built a supply-chain forecasting workflow using two fiscal years of sales history. Prophet generated the forecasts, and TabPy connected the Python model directly to Tableau so planners could view predictions beside the operational data they already used.

01Historical sales data
02Cleaning / aggregation
03Prophet model
04Monthly validation
05TabPy service
06Tableau decision layer

Connected Prophet to Tableau through TabPy instead of exporting static prediction files.

Measured forecast performance by month, with 94.25% average monthly accuracy across the project period.

Placed forecasts inside the existing Tableau workflow so users could compare model output with operational data.

Also worked with roughly 145,540 product-hierarchy records in related cleanup and analytics work, then presented findings to leadership.

The sales data needed hierarchy cleanup and consistent aggregation before it could be modeled reliably.

The forecast had to fit an existing business workflow, so integration and clear output mattered as much as model performance.

Reached 94.25% average monthly forecast accuracy and delivered the model through Tableau for operational use.

← All projectsHome →