# Using transform allows us to create a column which is the mean of a group within another column. # Say we want to add a column which is the mean of each strata, use the following code. measured_sg_a_strata_mean = sg_a_df.groupby([‘cea_strata’])[‘measured_carbon’].transform(‘mean’) sg_a_df[‘strata_mean_meas’] = measured_sg_a_strata_mean

## This gets n random samples from each group. # Create a sample DataFrame data = {‘group’: [‘A’, ‘A’, ‘B’, ‘B’, ‘C’, ‘C’, ‘D’, ‘D’], ‘value’: [1, 2, 3, 4, 5, 6, 7, 8]} df = pd.DataFrame(data) # Define a custom function to randomly select one value from each group def random_select(x): return x.sample(n=1).iloc[0] # Apply the function to each group in the DataFrame random_values = df.groupby(‘group’).apply(random_select)

Bootstrapping — is random sampling to estimate the standard deviation, or 95th confidence interval (or some other interval). Monte Carlo — is estimating the probability of an event. If one wants to get the mean of randomly sampling, I don’t think that is either monte carlo OR bootstrapping? ## Here is the code to get the mean and std of n random cores from 100 core samples. sample_number = 3 # choose n random samples from the group

# DataFrame to array ndarray = df.to_numpy() # Converting an array to a DataFrame: df = pd.DataFrame( X, columns = [‘col1’] )

### If I wanted to get the p value to list on the column of each graph ## steps — first partition the string## before, sep, p_CEA1 = str(stats.kruskal(cea_1_T0, cea_1_T1)).partition(‘pvalue=’) # then convert the second half of the text with ‘)’ in there to a 4 decimal place float. p_CEA1 = f’{float(p_CEA1.split(“)”)[0]):.4f}’ p_CEA1

To run non parametric stats (Kruskal Wallis, for n > 30), on a number of properties, use this code: ## not all data are normally distributed — use a Mann-Whitney or Kruskall Wallis # transform the data using SQRT transformation for p in range(0, len( property )): df = pd.read_excel(‘C:/Users/MelZeppel/OneDrive…

Different methods of normalizing data: import numpy as np a = np.random.rand(3,2) # Normalised [0,1] b = (a - np.min(a))/np.ptp(a) # Normalised [0,255] as integer: don't forget the parenthesis before astype(int) c = (255*(a - np.min(a))/np.ptp(a)).astype(int) # Normalised [-1,1] d = 2.*(a - np.min(a))/np.ptp(a)-1 Scaling using sklearn, including normalising the std dev. Write a function:

Merging is simple enough with two rows. However with multiple rows, concat is simpler. Remember to use [], and add axis = 1 to ensure columns are bound. df = pd.concat([cea_sample_97, cea_b_103, cea_model_only_90, cea_model_MV_sites_102], axis = 1)

Including 3 tables in the database. # Read in the Mullion FlintPro sqlite db con = sqlite3.connect(“Simplified_RothC_DB.sqlite”) cur = con.cursor() df = pd.read_sql_query(‘SELECT * FROM soil_inputs;’, con) # df = pd.read_sql_query(‘SELECT * FROM soil_inputs;’, con) # close the connection con.close() df # make a loop to make a dataframe (with…