Pandas - create dataframe manually and insert values

Providing a sample to increase your data frame dynamically... sizeOfDataFrame variable just limits for loop which adds data to the dataframe and is dynamic...

import pandas as pd
import numpy as np
yourDataFrame = pd.DataFrame()
sizeOfDataFrame = np.random.randint(100, size=1)
for currentLine in range(sizeOfDataFrame):
    yourDataFrame = yourDataFrame.append(pd.DataFrame({"A":np.random.randint(100, size=1),"B":np.random.randint(100, size=1),"C":np.random.randint(100, size=1)},index=[0]))
yourDataFrame.reset_index(inplace = True)    
yourDataFrame.drop('index',axis=1,inplace=True)

You can either initialize dataframe with data using

df = pd.DataFrame(columns=["A", "B"], data=[[5,np.nan]]),

or use set_value method (which is much faster than iloc by the way): df.set_value(0,'A',5)

UPDATE 2018-04-12 ⬇

Since pandas version 0.21.0 df.set_value is deprecated. You should use .at[]or .iat[] accessors instead:

df.at[0, 'A'] = 5