Assignment 8

a. Compute Combat Power (15 pts)

Again, the maximum combat power can be computed via information here for level 40 as:

Attack = 2 * round(attack^0.5 * sp_attack^0.5 + speed^0.5)
Defense = 2 * round(defense^0.5 * sp_defense^0.5 + speed^0.5)
Stamina = 2 * hp
MaxCP = (Attack + 15) * (Defense + 15)^0.5 * (Stamina + 15)^0.5 * 0.7903001^2 / 10

We wish to create a new column in our data frame named max_cp that stores this computed value. Do this without loops. Remember that pandas can do computations with series (columns).

b. Maximum Combat Power by Generation (10 pts)

Now, find the Pokémon with the greatest maximum combat power in each of the eight generations. Output the name of the pokemon along with it’s max_cp value.

a. Bar Chart (15 pts)

Using matplotlib directly or via pandas’ plotting routines, create a grouped bar chart that shows the mean attack, defense, and speed for each primary type. Set the figure size wider to improve the visualization. Include a legend.

b. Scatterplot (15 pts)

From the bar chart, we think there may be a negative correlation between attack and defense averages versus speed averages. First, compute a data frame that includes the mean attack, defense, and speed averages for each primary type, and then add the attack and defense means together to create a new column. Plot this new column versus the speed column. Label the axes appropriately. You should see one significant outlier with respect to the negative correlation. Determine the primary type of this outlier.

c. Interactive Scatter Matrix (25 pts)

Now, let’s use all three attributes and use altair to create a scatter matrix. A scatter matrix is a bunch of scatterplots, one for each pair of variables. Begin by creating a scatterplot that compares just attack and speed. Once you have this, change the x and y values to use the repeat capabilities of altair, and set the repeat over the speed, attack, and defense attributes. Finally, add a brush to the visualization so that we can interactively check relationships between the individual attributes. For example, try creating a rectangular selection over those primary types with high average attack and defense to see where their speeds land.

a. Bubble Chart (15 pts)

Using altair, create a bubble chart visualization that shows Pokémon positioned by attack and defense, sized by speed, and colored by generation. Make sure you aren’t biasing the visualization toward showing more from the later generations; this occurs by having later generations plotted on top of the older generations.

b. Binned Scatterplot (15 pts)

Since Part b led to major overdraw problems (points occluding other points), we will create a binned scatterplot that shows the count of the number of Pokémon in a particular attack/defense bin. This will help us better see the distribution of Pokémon according to attack and defense measures. Use 1200 bins (40x30) or some other reasonable number that shows the trend.