Trend Analysis of 80M Sales
Budget: $15 – $25 USD
I have a raw sales dataset that holds roughly eighty million records and I need a clear, defensible picture of the trends hidden inside it. The purpose is strictly insight-driven: I want to understand seasonality, product and regional performance shifts, and any emerging patterns that can guide strategic decisions.
You will start by examining the data quality, running the usual checks for duplicates, missing values, and outliers. Once it is clean, I expect you to apply the appropriate statistical and machine-learning techniques—time-series decomposition, clustering, cohort or basket analysis, whichever combination best surfaces trend signals. Python or R is fine (Pandas, NumPy, scikit-learn, tidyverse, etc.), and if you prefer a big-data stack such as PySpark, that works too; the volume will justify it.
Please package the outcome as:
• A concise written report (PDF or Markdown) that explains the key trends and how you arrived at them.
• Visualisations (static or interactive) that make the findings easy to consume for non-technical stakeholders—Matplotlib, Seaborn, Plotly, or Tableau Public dashboards are all acceptable.
• The cleaned dataset (or transformation scripts) plus fully commented code notebooks or scripts so I can reproduce the work end to end.
Accuracy, transparency, and reproducibility are critical. If that matches your skill set, I look forward to seeing how you can turn eighty million rows into actionable insight.
You will start by examining the data quality, running the usual checks for duplicates, missing values, and outliers. Once it is clean, I expect you to apply the appropriate statistical and machine-learning techniques—time-series decomposition, clustering, cohort or basket analysis, whichever combination best surfaces trend signals. Python or R is fine (Pandas, NumPy, scikit-learn, tidyverse, etc.), and if you prefer a big-data stack such as PySpark, that works too; the volume will justify it.
Please package the outcome as:
• A concise written report (PDF or Markdown) that explains the key trends and how you arrived at them.
• Visualisations (static or interactive) that make the findings easy to consume for non-technical stakeholders—Matplotlib, Seaborn, Plotly, or Tableau Public dashboards are all acceptable.
• The cleaned dataset (or transformation scripts) plus fully commented code notebooks or scripts so I can reproduce the work end to end.
Accuracy, transparency, and reproducibility are critical. If that matches your skill set, I look forward to seeing how you can turn eighty million rows into actionable insight.
Related categories:
Python
Statistics
Machine Learning (ML)
Statistical Analysis
Data Science
Data Visualization
Data Analysis
Pandas