Assignment – Data Science Associate
Objective:
The objective of this assignment is to analyse a dataset regarding the number of
goals scored in soccer matches. Candidates are expected to apply statistical
methods to derive insights and fit appropriate probability distributions.
Dataset:
You are provided with data containing the number of goals scored in 2000 soccer
matches. The dataset includes the following information:
Total number of matches = 2000
Number of goals: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9
Number of matches: 150, 400, 600, 500, 250, 75, 15, 4, 1, 5
Tasks:
• 1. Calculate the probabilities for each value of X (number of goals) based on
the provided data.
• 2. Compare the observed probabilities with those predicted by the Poisson
distribution. Estimate the parameter λ from the data as the average number
of goals per match.
• 3. Fit the Negative Binomial distribution to the data and determine its
parameters (r and p). Justify your choice of distribution based on the shape of
the probability distribution.
• 4. Create visualisations to represent the observed frequencies and fitted
distributions using a Python library (e.g., Matplotlib or Seaborn).
• 5. Write a brief report summarising your findings, including:
- The calculated probabilities
- The comparison with the Poisson distribution
- The results of the Negative Binomial fit
- Any conclusions and recommendations based on your analysis.
Coding Requirements:
1. Use Python for your analysis and provide well-commented code to ensure
readability.
2. Submit your code in a Jupyter Notebook(.ipynb) or as a Python script(.py).
Question 2
1. Exploratory Data Analysis (EDA)
• Task: Perform EDA on the leads dataset. Include visualizations and
descriptive statistics to analyze key aspects of the data.
2. Lead Scoring Logic Development
• Task: Develop a scoring model to identify leads with a higher probability of
conversion.
o Considerations:
▪ Use relevant features from the dataset to calculate a lead
score.
▪ Define a scoring formula or logic based on observed patterns
and insights gained from EDA.
o Questions to Address:
▪ What factors significantly contribute to a higher lead score?
▪ How will you test and validate the scoring model for
effectiveness?
3. Top Contributors/Features
• Task: Identify and rank features that contribute most significantly to lead
conversion.
o Questions to Address:
▪ Which features (e.g., UTM-Source, CountOfClickEvents,
WebTimeSpent) have the highest impact on successful
conversions?
▪ How can these insights be used to prioritize leads and
optimize efforts?
4. Differentiation Features for Positive and Negative Sale Results
• Task: Analyze features that differentiate successful leads from those that
did not convert.
o Questions to Address:
▪ What are the common characteristics of leads that
converted successfully compared to those that did not?
▪ Are there specific behavioral patterns that correlate with
positive or negative outcomes?
▪ Can you propose any recommendations for improving lead
conversion based on your findings?
Coding Requirements:
1. Use Python for your analysis and provide well-commented code to ensure
readability.
2. Submit your code in a Jupyter Notebook(.ipynb) or as a Python script(.py).
Deliverables
• A report summarizing findings from the EDA, including visualizations and
key insights.
• A lead scoring model and rationale for the scoring logic used.
• A presentation highlighting the top contributors/features and
differentiation features for sale results.