我可以将值分成桶吗？

Question

我有一个包含名称和购买价值的两列数据框。有什么方法可以将它们聚集在 25% 的桶中，以显示它们在总价值中的占比。

因此，前 25% 的销售额占总收入的 50%

我已经按值从小到大进行了分组和排序

df = df.sort_values(['purchase_value'], ascending=[True])

我想要一个表格输出显示：

集团总收入百分比

前 25% - 56%

第二个 25% - 25%

降低 25% - 15%

底部 25% - 4%

Answer 1

numpy 提供了一些方便的统计功能，您可以使用这些功能来实现您所需要的。特别是numpy.percentile 和numpy.histogram。

import numpy as np

# 100 random numbers for testing purposes
values = np.random.rand(100)
# values = np.array(df.purchase_value)

# Get the values on the boundary of each percentile
q_list = [0, 25, 50, 75, 100]
percentiles = list(np.percentile(values, q_list))

# Sum the values in the bin between each pair of percentiles
bin_sums = np.histogram(values, bins=percentiles, weights=values)[0]

# Normalize by the total amount of value
bin_percents = 100 * bin_sums / np.sum(bin_sums)

# Output
for left, right, percent in zip(q_list[:-1], q_list[1:], bin_percents):
    print("Portion of value between {}% and {}%: {}%".format(left, right, percent))

# Portion of value between 0% and 25%: 6.7259763107428965%
# Portion of value between 25% and 50%: 17.0792197951553%
# Portion of value between 50% and 75%: 31.52807116533235%
# Portion of value between 75% and 100%: 44.66673272876946%

---