'myEdu': 'myEdu'
}
# Collect info about each level
final_nos['levels'] = nos_levels[0][0]
group_levels = final_nos.groupby('levels')
# edu/exp
nos_groups = group_levels[['myExp-peak', 'myEdu-peak']]
nos_groups = nos_groups.agg(
lambda x: Counter(x).most_common()[0][0]) #np.max)
# average salary
nos_groups2 = group_levels['Salary-peak']
nos_groups2 = nos_groups2.agg(np.median).map(np.round)
nos_groups = nos_groups.join(nos_groups2)
# NOS titles
final_nos['NOS Title'] = final_nos['NOS Title'].map(
lambda x: x.capitalize())
nos_groups2 = group_levels['NOS Title']
nos_groups2 = nos_groups2.apply('\n'.join)
nos_groups = nos_groups.join(nos_groups2)
# URNs
nos_groups2 = group_levels['URN']
nos_groups2 = nos_groups2.apply('\n'.join)
nos_groups = nos_groups.join(nos_groups2)
# top 10 skills
nos_groups2 = pd.DataFrame(
group_levels['converted_skills'].agg('sum').map(
lambda x: x.most_common()).map(lambda x: x[:20]).map(
lambda x: [t[0].capitalize() for t in x]).map(
'\n'.join))
# mark skills that are not "public"
#tmp = pd.DataFrame(group_levels['converted_skills'].agg(
