fixture difficulty tool

[kad568]

what makes a team hard to play?

• good defence
• good attack

[kad568]

old fpl json has defensive and offensive ratings

[kad568]

use seaborn for plotting heatmaps

[kad568]

[kad568]

maybe sort by avg heat rating

• also check normalisation method
  • use min and max of the entire data set not each colomn

[kad568]

To rank a Seaborn heatmap by average heat, you want to reorder the rows or columns based on the average value of the heatmap's data. This involves calculating the average for each row or column and then sorting accordingly. Here's a general approach using Python and Seaborn:

1. Calculate Averages: Compute the average value for each row or column of your heatmap data.
2. Sort Data: Sort the rows or columns based on these averages.
3. Reorder Heatmap: Use the sorted indices to reorder your data in the heatmap.

Here's a step-by-step code example:

import seaborn as sns
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

# Sample data for illustration
data = np.random.rand(10, 10)  # 10x10 matrix with random values
df = pd.DataFrame(data, columns=[f'Col{i}' for i in range(1, 11)], index=[f'Row{i}' for i in range(1, 11)])

# Calculate the average heat for each row
row_averages = df.mean(axis=1)

# Get the order of rows based on the average heat (from highest to lowest)
row_order = row_averages.sort_values(ascending=False).index

# Reorder the DataFrame based on this order
df_ordered = df.loc[row_order]

# Create the heatmap with reordered rows
sns.heatmap(df_ordered, cmap='viridis', annot=True)
plt.title('Heatmap Ordered by Average Heat')
plt.show()

Explanation:

• Data Preparation: We create a DataFrame df representing the data for the heatmap.
• Calculate Averages: We calculate the average for each row using df.mean(axis=1).
• Sort and Reorder: We sort these averages and reorder the DataFrame df accordingly.
• Plotting: Finally, we plot the heatmap with the reordered data.

Note: If you want to rank by column averages instead, switch axis=1 to axis=0 when calculating the averages and sort the columns instead of the rows.

[kad568]

i need a stat to compare performance, so i can corelate team strength effects on opposition:

• team strength difference
• compare to difference to average performance

performance could mean xg made, xg conceded, xg difference, g for and against difference

do a scatter plot to get an average effect strength differences has so 50% decrease when vs city for example

[kad568]

imrpoved

[kad568]

weird, the away strength is high in all cases, which does not match intuition

[kad568]

• could confirm this with the xg difference data
• adjust the scores acordingly

[kad568]

maybe an issue on my end

[kad568]

the actual team starts are in json files hosted online, see:

• https://data.fotmob.com/stats/47/season/20720/goals_team_match.json

[kad568]

maybe the same for player.

use these links instead

[kad568]

got data