In the previous section, we discussed the calculation of Series and the automatic alignment of Pandas. Not only Series, DataFrame also supports computation and is one of the most commonly used features.
Because DataFrame data structures contain multiple rows and columns, DataFrame calculations and statistics can be performed using either row data or column data. For our convenience, Pandas provides common computational and statistical methods:
| operation | methods | operation | methods |
|---|---|---|---|
| sum | sum | The maximum | max |
| Calculating mean | mean | The minimum value | min |
| Strives for the variance | var | The standard deviation | std |
| The median | median | The number of | mode |
| quantile | quantile |
Operation.
We have N students’ scores in math, Chinese and English. Now, we want to calculate the total score of each student. We can use the following method:
"' the sum demonstrates two methods: method 1: the column data to sum first delete selected (remove the name column), then use the sum function sum method 2: Select the columns to be summated one by one and use the operator to sum the two methods. The result is the sum of each row in the DataFrame.
df['sum'] = df[['chinese'.'math'.'english']].sum(1) Method # 1
df['sum'] = df['chinese'] + df['math'] + df['english'] Method # 2
Output：
name chinese english math sum
0 XiaoMing 99 100 80 279
1 LiHua 102 79 92 273
2 HanMeiNei 111 130 104 345
Copy the codeIn the sum method we pass in the argument 1, which represents the row we are using. To calculate the sum for each column, we simply pass in 0:
df[['chinese'.'math'.'english']].sum(0) the Output: Chinese312
math 276
english 309
dtype: int64
Copy the codeNow that we have the total score, the math teacher or the Chinese teacher will care about the average score of the students in the class, and again, we can calculate it very quickly:
df['math'].mean() # Method 1: Use the mean method provided by Pandas
df['math'].sum() / df.shape[0] # method 2: Use summation method to calculate the sum and divide by the total number of rowsThe Output:92.0
Copy the codeThe DataFrame shape method is used in 🌰. This method displays the number of rows and columns of the DataFrame. The number of rows is 0 and the number of columns is 1. Note that the output column values do not include indexed columns.
The above 🌰 only calculated the average score of math, interested partners can base on their own English and Chinese average score oh ~
2. Statistical
I’ll look at the number of math scores for the students in their class. Pandas have the lowest and highest scores in math.
df['math'].min(a)The minimum value of the math columnThe Output:80
df['math'].max(a)The maximum value of the math columnThe Output:104
df['math'].quantile([0.3.0.4.0.5]) The 30%, 40%, 50% quantile of the math columnThe Output:0.3 87.2
0.4 89.6
0.5 92.0
Name: math, dtype: float64
df['math'].std() The standard deviation of the math columnThe Output:12
df['math'].var() The variance of the math columnThe Output:144
df['math'].mean() The average of the math columnThe Output:92
df['math'].median() The median of the math columnThe Output:92
df['math'].mode() The mode of the # math column, which returns a Series object (parallelism is possible, in this case the mode is 1, so return both)
Output:
0 80
1 92
2 104
dtype: int64
Copy the codeWe can also use the Describe method of the DataFrame to view basic statistics on the DataFrame:
df.describe()
Outprint:
chinese english math sum
count 3.000000 3.000000 3.0 3.000000
mean 104.000000 103.000000 92.0 299.000000
std 6.244998 25.632011 12.0 39.949969
min 99.000000 79.000000 80.0 273.000000
25% 100.500000 89.500000 86.0 276.000000
50% 102.000000 100.000000 92.0 279.000000
75% 106.500000 115.000000 98.0 312.000000
max 111.000000 130.000000 104.0 345.000000
Copy the code3. Sorting
Generally speaking, our score table is in order of total points from highest to lowest:
df = df.sort_values(by='sum', ascending=False)
Output:
name chinese english math sum
2 HanMeiNei 111 130 104 345
0 XiaoMing 99 100 80 279
1 LiHua 102 79 92 273
Copy the codeYou can see that we use the sort_values method to sort the DataFrame, and the by argument ‘sum’ is passed to sort according to the ‘sum’ field. Ascending sets whether the order is descending (False) or ascending (True, the default). Sort_values returns a new DataFrame by default, which does not affect the original DataFrame. In this example, we assign the sorted DataFrame to the original DataFrame. If we do not want to create a new DataFrame, we can do so. Just pass inplace=True (to modify the original DataFrame) :
df.sort_values(by='sum', ascending=False, inplace=True)
print(df)
Output:
name chinese english math sum
2 HanMeiNei 111 130 104 345
0 XiaoMing 99 100 80 279
1 LiHua 102 79 92 273
Copy the codeCareful friend may find when we sort, if you have any adjustment DataFrame of row data, the adults of the index value is not changed, the above example because we use the default index of ascending series, so after ordering does not look very friendly, but don’t worry, we can reset the index values of:
df = df.sort_values(by='sum', ascending=False).reset_index()
Output:
index name chinese english math sum
0 2 HanMeiNei 111 130 104 345
1 0 XiaoMing 99 100 80 279
2 1 LiHua 102 79 92 273
Copy the codeAfter resetting the index with reset_index, the index column of our DataFrame object is indeed reset to an increasing sequence, as well as an additional column named index. Of course we can pass drop=True to insert the original index column into the new DataFrame:
df = df.sort_values(by='sum', ascending=False).reset_index(drop=True)
name chinese english math sum
0 HanMeiNei 111 130 104 345
1 XiaoMing 99 100 80 279
2 LiHua 102 79 92 273
Copy the codeTo visualize the ranking, we can index +1 to show the ranking of students:
df.index += 1
name chinese english math sum
1 HanMeiNei 111 130 104 345
2 XiaoMing 99 100 80 279
3 LiHua 102 79 92 273
Copy the code