dusteenie / date-a-scientist Goto Github PK

Rubric Score

Criteria 1: Valid Python Code

• Score Level: 3 (Meets expectations)
• Comment(s): Your code runs with a small error (see issue #2), but it was easy to fix.

Criteria 2: Exploration of Data

• Score Level: 2 (Approaches expectations)
• Comment(s): Your data exploration is off to a good start, and you did a good job thinking about how to convert the text values to numbers. Instead of choosing values ranging from 0 to 1, I recommend splitting the values up into two categories, and assigning all values in one category to a 1 and the values in another category to a 0 (e.g., smokes = 1, doesn't smoke = 0...has college education = 1, doesn't have college education = 0...etc.). This makes both the running and interpreting of the model easier, and is a more common practice than what you chose to do. You can also leave the continuous values like height and income as-is. Also, in general your data exploration should inform your research question to a greater extent. Try making some histograms of the distributions of different variables, and make a scatter plot or calculate correlations to look at the relationship between variables.

Criteria 3: Machine Learning Techniques used correctly

• Score Level: 4 (Exceeds expectations)
• Comment(s): Good job using regression models to predict continuous outcomes and classification models to predict categorical outcomes. Also, nice job comparing the models.

Criteria 4: Report: Are conclusions clear and supported by data?

• Score Level: 2 (Approaches expectations)
• Comment(s): Nice job discussing your results, but when you analyze your regression model results, instead of accuracy and precision what you're actually displaying are the R^2 score for the training and test sets. R^2 is the measure that we use to analyze a regression model's performance, and it represents the proportion of variance explained by the model. Also, in your classification model results, the graphs you display are not very informative. It would be better if they were scatter plots (without the lines), but even then you would just have a point at each of the four corners. Finally, it would be better to mention F1 score than to focus on precision -- your F1 scores are around 50% (chance), and are a better measure of model performance than precision or accuracy (although those scores are useful too, good job including them).

Criteria 5: Code formatting

• Score Level: 4 (Exceeds expectations)
• Comment(s): Good job, your code was well-formatted. Good use of comments to break your code up into sections.

Overall Score: 15/20

These lines of code gives me an error, because plt does not have a score. When I changed plt to rr, your code ran fine.

Is this value supposed to be 1 instead of 8?