End to end retail sales analysis and forecasting project built using R for the data preparation and modeling and Power BI for the interactive dashboard and insights.
This project focuses on taking raw retail transaction data, turning it into a clean analytical dataset, building clear KPIs, exploring what drives sales and profit, and producing simple, realistic forecasts based on historical trends.
I built this project to apply what I learned after completing the Microsoft Power BI Data Analyst Professional Certificate (Coursera), and to combine those skills with my background in R and data analysis.
Starting from a raw global retail dataset, I cleaned and transformed the data in R, created Power BI ready tables, and designed a multi page dashboard that highlights performance over time, differences across categories and regions, product level behaviour, and future outlook.
The final result is an interactive Power BI dashboard supported by a reproducible data preparation pipeline.
- Dataset: Global Superstore Dataset
- Source: https://www.kaggle.com/datasets/apoorvaappz/global-super-store-dataset
- Time Period: 2011 to 2014
- Scope: Global retail transactions across multiple markets, regions, product categories, and customer segments
Key fields include:
- Order dates and order identifiers
- Sales, profit, quantity, discounts, and shipping costs
- Product, category, and sub category information
- Geographic attributes such as market, region, and country
The dataset works well for both descriptive analysis and time series based forecasting.
I built this project to:
- Practice structuring a full analytics project from raw data to final dashboard
- Strengthen my skills in R for data cleaning and feature engineering
- Apply Power BI for data modeling, DAX calculations, and dashboard design
- Include forecasting rather than only historical analysis
- Build a project that reflects practical analytics work
A big focus throughout the project was on clear data visualization. The goal was not to add complexity, but to present information in a way that is easy to understand and useful for decision making.
The project followed a clear and structured process:
- Loaded the raw Global Superstore dataset into R
- Cleaned and standardized column names and data types
- Parsed and engineered date features such as year, month, and year month
- Created clean, analysis ready tables for Power BI
- Saved processed datasets to a structured
data/processedfolder - Designed a fact table focused on sales transactions
- Imported processed data into Power BI
- Built relationships and validated the data model
- Created core KPIs such as total sales, total profit, total orders, and average order value
- Developed DAX measures including profit margin
- Designed multi page dashboards with consistent layout and filtering
- Analysed category, region, and product level performance
- Applied Power BI forecasting to monthly sales and profit
- Interpreted forecast results carefully and realistically
- Summarised key insights, risks, and next steps in an executive style page
All data preparation steps are fully reproducible using the R scripts in this repository.
The Power BI dashboard consists of five focused pages, each answering a specific set of questions.
High level KPIs and time series trends showing overall sales, profit, order volume, and growth over time. Interactive slicers allow filtering by year, market, and customer segment.
Breakdown of sales and profit by product category and geographic region. This page highlights clear performance differences across markets and categories.
A closer look at product and sub category performance, showing which products drive revenue versus profit and where margin pressure exists.
Twelve month forecasts for monthly sales and profit based on historical patterns from 2011 to 2014. The forecasts are intended as directional guidance rather than exact predictions.
A concise summary of key insights, risks, and suggested next actions.
- Technology products bring in a large share of total sales, but their profit margins are less consistent compared to other categories.
- Office Supplies tend to generate smaller sales per order, but profits are more stable across regions.
- Higher discounts are often linked to lower profit, especially for products that already have tight margins.
- Some regions show strong sales growth but contribute less to overall profit, suggesting efficiency differences across markets.
- The sales forecast suggests that overall sales are likely to keep increasing over the next year, following the upward trend seen in the historical data.
- Profit is also expected to grow, but more slowly and with noticeable fluctuations, which may point to ongoing margin pressure.
These forecasts are meant to give a general idea of where performance could be heading, rather than exact future values.
- Take a closer look at discounting on high volume products that are not contributing much to profit.
- Focus on growing categories and regions that show both strong sales and reasonable margins.
- Revisit and update the forecasts regularly as new data becomes available.
Due to Power BI online sharing limitations for personal accounts, the interactive dashboard is shared as a PBIX file.
The full dashboard can be downloaded here and explored locally in Power BI Desktop: https://1drv.ms/u/c/77409db4c05f6c66/IQAhh4s8EnPmTYXkmePLV0LyAUFvHxMrGzgxQF-We66FvdY?e=MsDVQT
All five dashboard pages are also documented using screenshots above to provide a clear preview of the layout, visuals, and insights. To view the dashboard:
- Download the PBIX file from the link above
- Open it using Power BI Desktop
- Interact with filters, slicers, and visuals locally
Forecasts were created using Power BI’s built in time series forecasting on monthly aggregated data.
Key observations:
- Sales are expected to continue following the existing upward trend
- Profit growth appears slower than sales growth
- Seasonal patterns remain visible in the forecast
- Confidence intervals are included to reflect uncertainty
The forecasts are positioned as support for planning rather than precise predictions.
- R
- tidyverse
- Power BI
- Power Query for data transformation
- DAX for calculated measures and KPIs
- Data cleaning and transformation
- Data modeling and relationships
- KPI design and validation
- Time series analysis and forecasting
- Data visualization and dashboard design
- Communicating insights clearly
-
data/processed/
Cleaned and transformed datasets prepared in R and used directly in Power BI. -
scripts/
R scripts used for data cleaning, feature engineering, and preparing Power BI ready tables. -
powerbi/screenshots/
Screenshots of each Power BI dashboard page. -
Retail-Sales-Insights-and-Forecasting.pbix
Power BI dashboard file. -
README.md
Project documentation. -
.gitignore
Git configuration file.
The full interactive dashboard can be explored locally by opening the Power BI file.
This project focuses on clarity and solid fundamentals rather than visual overload. The dashboard is designed to be easy to read, interactive, and useful for exploring the data.
If you would like to connect or give feedback on this project (much appreciated):