Business Dashboard

A Streamlit dashboard that surfaces trends, forecasts demand with Prophet, and flags anomalies so teams act faster.

Tools: Python • Streamlit • Prophet • Plotly • Pandas Role: Solo builder Users: Friends and early testers with real reporting needs

Problem

Small teams had sales and profitability data but no simple way to spot trends, forecast, or catch anomalies. Reports were manual and late, so decisions lagged.

Solution

I built a Streamlit dashboard with Prophet forecasting, anomaly detection, and Plotly visuals. Data prep used Pandas with configurable filters and date-grouping. UX prioritized speed to insight: KPIs up top, trends and drill-downs below.

Impact

Test users reduced weekly reporting time from hours to minutes and surfaced seasonal patterns they hadn’t seen. The anomaly view helped flag outliers that led to pricing and inventory fixes.

Role & Context

Solo project in ~2 weeks. End users were friends running small operations; I iterated based on their feedback. Tech: Python, Streamlit, Prophet, Plotly, Pandas.

KEY FEATURES

  • Time-Series Forecasting: Uses Facebook Prophet for demand forecasting with configurable parameters
  • Anomaly Detection: Automatically identifies outliers and unusual patterns in sales data
  • Interactive Visualizations: Plotly charts for trend analysis, forecasting, and anomaly visualization
  • Data Filtering: Configurable date ranges and grouping options for flexible analysis
  • KPI Dashboard: Key metrics displayed prominently for quick insights
  • Export Capabilities: Export forecasts and analysis results

SKILLS DEMONSTRATED

Data Cleaning Time Series Analysis Forecasting Anomaly Detection Data Visualization Streamlit Development Python Pandas Plotly Prophet

TECHNOLOGIES USED

Python Streamlit Prophet Plotly Pandas NumPy

Demo

CHALLENGES & SOLUTIONS

Challenge: Time-series data required careful preprocessing and handling of missing values
Solution: Implemented robust data cleaning pipelines with Pandas, handling missing values and ensuring consistent date formats before feeding into Prophet.
Challenge: Prophet model parameters needed tuning for accurate forecasts
Solution: Used cross-validation and experimentation to find optimal Prophet parameters, balancing forecast accuracy with computational efficiency.
Challenge: Creating intuitive visualizations for both technical and non-technical users
Solution: Designed clear Plotly charts with interactive tooltips and clear labels, making complex forecasts and anomalies accessible to all users.

Explore Related Projects

Interested in more data analytics projects? Check out these related dashboards:

Expense Tracker Dashboard Loan Portfolio Dashboard Online Retail Dashboard View All Projects