Section Description Points Grade Breakdown and Requirements Weightage

60 What 80-100% looks like What 60-80% looks like What <60% looks like 100.00%
Exploratory Data Analysis - Problem definition, questions to be 12 1) Definition of problem [1] 1) Definition of problem 1) Definition of problem 20.00%
answered
- Data background and contents 2) Observations on shape of data, data types of 2) Observations on shape of data, data types of 2) Observations on shape of data and data types
- Univariate analysis various attributes, statistical summary. [2] various attributes, statistical summary. of various attributes
- Bivariate analysis
- Answers to the key questions 3) Insight-based questions given in learner 3) Insight-based questions given in learner 3) Some of the insight-based questions given in
provided notebook answered. [6] notebook answered. learner notebook answered.
- Insights based on EDA
4) Univariate and bivariate analysis (variable
distributions, interactions between variables) to
understand the relationships in data beyond the
set of questions already provided. [3]

* Marks for the key questions provided are

included within the univariate and bivariate
analysis marks
Data preprocessing - Duplicate value check 8 1) Checked for duplicate values [2] 1) Checked for missing values 1) Checked for missing values 13.33%
- Missing value treatment
- Outlier check (treatment if needed) 2) Checked for missing values [3] 2) Checked for outliers
- Feature engineering
- Data preparation for modeling 3) Checked for outliers [3] * Outlier treatment is not mandatory and no
points to be deducted for not performing it
* Outlier treatment is not mandatory and no
points to be deducted for not performing it
** Feature engineering and transformations to
reduce skewness in data is good to have (no
points to be deducted for this)
Model building - Linear - Build the model and comment on the 10 1) Linear regression model built [5] 1) Linear regression model built 1) Linear regression model built 16.67%
[email protected]
Regression model statistics
QN3DYCG1XF - Display model coefficients with 2) Model performance - R2 >= 0.7 [2] 2) Model performance - R2 between 0.65 and 0.7 2) Model performance - R2 < 0.65
column names
3) List of model coefficients with column names 3) List of model coefficients with column names 3) Train and test R2 differ more than 0.05
[2]
4) Train and test R2 differ less than 0.05
4) Train and test R2 differ less than 0.05 [1]
Testing the assumptions of linear - Perform tests for the assumptions of 12 All the test mentioned below are performed and Three of the tests are performed and conditions Any test is performed 20.00%
regression model the linear regression conditions are satisfied are satisfied
- Comment on the findings from the
tests 1) Multicollinearity check by VIF score (variables
are dropped one-by-one till none has VIF>5) [4]

2) Linearity of variables (no pattern in residual

plot) [1.5]

3) Independence of residuals (no pattern in

residual plot) [1.5]

4) Normality of residuals (almost bell-shaped

curve in residuals distribution, points in QQ plot
are almost all on the line) [3]

5) Test for Homoscedasticity (p-value > 0.05) [2]

Model performance evaluation Evaluate the model on different 6 1) Metrics checked - MAE, RMSE, R2, Adj R2 [3] 1) Metrics checked - MAE, RMSE, R2 1) Any one of MAE, RMSE, or R2 is checked 10.00%
performance metrics
2) Train and test performances are checked [2] 2) Train and test performances are checked

3) Comments on the performance measures and if

there is any need to improve the model or not [1]

This file is meant for personal use by [email protected] only.

Sharing or publishing the contents in part or full is liable for legal action.
 Actionable Insights & - Comments on significance of 6 1) 3-4 Actionable Insights mentioned, including 1) 1-2 Actionable Insights mentioned, including Any Actionable Insights and Recommendations 10.00%
Recommendations predictors comments on significance of predictor variables comment on significance of predictor variables mentioned
- Key takeaways for the business [4]
2) 1 Recommendation mentioned
2) 2-3 Recommendations mentioned [2]

[Recommendations can also include points on

additional data sources for further analysis, model
implementation in real world, potential business
benefits from improving the model, etc.]
Quality of Business Report 6 - Objective, guidance, and data description: 1 - Objective, guidance,and data description - Objective, guidance,and data description 10.00%
point - Couple of lines of code included - Multiple lines of code included
- Exclusion of code: 2 points - Structure and readability - Lak of structure and readability (non-sequential
- Structure and readability: 1 point - Rationale and logic anwering of questions)
- Rationale and logic: 1 point - Plots and tables do not have visual clarity and - Rationale and logic
- Visual clarity and referencing: 1 point there is no referencing - Plots and tables do not have visual clarity and
there is no referencing

[email protected]
QN3DYCG1XF

This file is meant for personal use by [email protected] only.

Sharing or publishing the contents in part or full is liable for legal action.