Datathon Summary: Analysis and Challenges

Overview

During the datathon, our team focused on analyzing workforce dynamics and customer behavior in a business setting. We tackled three key problems:

Analyzing the Relationship Between Profit and Number of Workers

• We explored how staffing levels influence profitability by correlating financial performance with workforce size.
• Our approach involved regression analysis and trend identification to determine optimal staffing levels.

Predicting Number of Employees per Hour Based on Orders per Hour

• We developed a predictive model to determine the required workforce at different times of the day based on order volumes.
• Machine learning techniques, such as linear regression and time series analysis, were employed to optimize workforce allocation.

Understanding Tipping Percentage vs. Hour of Day

• We analyzed tipping trends to understand customer behavior throughout the day.
• Insights were gathered on how factors like peak hours and customer demographics impact tipping patterns.

Challenges Faced

Data Quality and Cleaning

• The dataset contained inconsistencies, missing values, and noise, requiring extensive preprocessing.

Feature Selection and Model Optimization

• Determining the most relevant features for prediction proved challenging, necessitating iterative model refinement.

Time Constraints

• Given the limited timeframe, balancing exploratory data analysis, model development, and result interpretation was difficult.

Computational Limitations

• Running complex models on large datasets required significant computing power, leading to occasional slowdowns.

Interpreting Business Implications

• Translating data-driven insights into actionable business strategies posed a challenge, requiring domain knowledge and logical reasoning.

End Goal and Conclusion

Our ultimate goal was to develop a model that could efficiently schedule shifts based on demand, ensuring optimal workforce allocation while maximizing profitability. Despite the challenges, our team successfully built models that provided valuable insights into workforce optimization, profitability, and customer behavior. The datathon was an enriching experience, improving our data analysis, machine learning, and problem-solving skills.

Built With

• pandas
• sckitlearn