Something like this, to begin with:

min max mean    std count   count %  
factor_quantile  
1   -0.999794   0.227477    -0.253214   0.178138    922640  20.019381  
2   -0.407817   0.406284    -0.015098   0.113215    921293  19.990153  
3   -0.266352   0.611015    0.110669    0.112861    921294  19.990175  
4   -0.141869   0.946058    0.247604    0.131643    921292  19.990132  
5   -0.005262   33.430894   0.677265    0.652337    922215  20.010159  

So here's a little update, am I doing right?

The Number of elements column is not right, instead of

([2], [1.222, 2.6162460000000003]) I would like to get just
2

Good question! First, add a column to the dataframe which indicates which quintile each row belongs to

df['quintile'] = pd.qcut(df.factor, q=5, labels=[1, 2, 3, 4, 5])

Next, group on the quintile and use the aggregate or agg method to apply any calculations to each group. Put the names of the functions to apply in a list.

df_stats = df.groupby('quintile').factor.agg(['min', 'max', 'mean', 'std', 'count'])

The 'count %' function is a little special. It relies on knowing the sum of all the counts to get a percentage. The group functions only know about the factor values in their group and not the total. The simplest way to get the percent is so do this as a separate step. Note that one must use bracket, and not dot, notation because 'count' is also a method name.

df_stats['count %'] = df_stats['count']/df_stats['count'].sum()

See the attached notebook. Hope that's what you were looking for.

Dan, you always amaze me!
Thank you

sorry to disturb you again for a very noob question, can't we put the code in a function so that I can pass a dataframe and a column and it works?
Something like:

def quintile(df,column):  
    df['quintile'] = pd.qcut(df.column, q=5, labels=[1, 2, 3, 4, 5])  
    # First apply the functions which only rely on the groups  
    df_stats = df.groupby('quintile').column.agg(['min', 'max', 'mean', 'std', 'count'])  
    # The 'count %' stat relies on knowing the sum of all the counts so do this as a separate step  
    # Note that one must use bracket, and not dot, notation because 'count' is also a method name  
    df_stats['count %'] = df_stats['count']/df_stats['count'].sum()  
    df_stats  

if I have a dataframe called "df5" and it has a column with all numbers such as factor called "factor3" I would like to run
quintile(df5,factor3)

and have just this printed:
```

min max mean    std count   count %  

quintile
1 1.22200 2.44000 1.83100 0.861256 2 0.285714
2 3.32123 3.32123 3.32123 NaN 1 0.142857
3 6.88800 6.88800 6.88800 NaN 1 0.142857
4 7.84423 7.84423 7.84423 NaN 1 0.142857
5 13.00000 19.00000 16.00000 4.242641 2 0.285714
```

Always glad to help out @ Emiliano. It's good to know the pandas methods. Even if one doesn't remember exactly how they work, just knowing that things like groupby and agg exist is a start. I often can't recall the specifics so I google something like 'how to groupby with multiple functions in pandas 0.18'. My usual 'go to' links in the results are the pandas docs and stackoverflow. I add '0.18' in the search to get the docs for pandas version 0.18 which Quantopian currently uses. It's frustrating to find a solution only to realize it doesn't work because it relies on a more current version of pandas.

Also, I just noticed you asked about plotting. Not sure what you want to plot, but I often look at the distribution of factor values using the hist method. It really jumps out if the factor values are not evenly distributed or there are big outliers.