def user_pull():
"""
Pull app creators Mixpanel information.
Parameters
----------
Global Variables
----------
api_creator_secret: Mixpanel Creator Project API key.
Use to make API calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe containing app creator information from Mixpanel.
"""
#generate JQL query
query_user = JQL(api_creator_secret,
people=People({'user_selectors': [{}]
})).group_by(keys=[
"e.properties.$email",
"e.properties.$username",
"e.properties.$distinct_id",
"e.properties.hs_mrr"
],
accumulator=Reducer.count())
#initialize list to track emails, user IDs, distinct IDs, and mrr
email_list = []
user_id_list = []
distinct_id_list = []
hs_mrr_list = []
#process query results
for row in query_user.send():
email_list.append(row['key'][0])
user_id_list.append(row['key'][1])
distinct_id_list.append(row['key'][2])
hs_mrr_list.append(row['key'][3])
#create dataframe
data = {
'email': email_list,
'user_id': user_id_list,
'distinct_id': distinct_id_list,
'hs_mrr': hs_mrr_list
}
df_users = pd.DataFrame(data=data)
return df_users
def AppStart_pull(from_date, to_date):
"""
Pull app user AppStart events in Mixpanel.
Parameters
----------
from_date: date
Start date of query.
to_date: date
End date of query.
Global Variables
----------
api_user_secret: str
Client secret used to make calls to Mixpanel User Project.
Returns
----------
dataframe
Dataframe contains user IDs and AppStart event count for app users in Mixpanel.
"""
#generate JQL query
query = JQL(api_user_secret,
events=Events({
'event_selectors': [{
'event': "AppStart"
}],
'from_date': from_date,
'to_date': to_date
})).group_by(keys=["e.properties.zUserId"],
accumulator=Reducer.count())
#initalize lists to record user ID and AppStarts
user_id_list = []
AppStart_list = []
#process query results
for row in query.send():
if row['key'][0] is not None:
user_id_list.append(int(row['key'][0]))
AppStart_list.append(row['value'])
#generate dataframe
data = {'user_id': user_id_list, 'AppStart': AppStart_list}
df_AppStart = pd.DataFrame(data)
df_AppStart = df_AppStart.dropna()
df_AppStart.user_id = df_AppStart.user_id.astype(int)
return df_AppStart
def get_new_signup(yesterday):
"""
Get yesterday new signup.
Parameters
----------
yesterday: date
Yesterday's date.
Global Variables
----------
api_creator_secret: Mixpanel Creator Project API key.
Use to make API calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe containing user IDs and emails of creators who signed up yesterday.
"""
# New Signup Web
query = JQL(api_creator_secret,
events=Events({
'event_selectors': [{
'event': "New Signup Web"
}],
'from_date': yesterday,
'to_date': yesterday
}),
people=People({'user_selectors': []}),
join_params={
'type': 'full',
'selectors': [{
'event': "New Signup Web"
}]
})
#store email, user id, and sign up events
email_list = []
userid_list = []
for row in query.send():
if 'user' in list(row.keys()):
if '$username' in list(row['user']['properties'].keys()):
userid_list.append(row['user']['properties']['$username'])
if '$email' in list(row['user']['properties'].keys()):
email_list.append(row['user']['properties']['$email'])
#create dataframe
data = {'user_id': userid_list, 'email': email_list}
df_new_users = pd.DataFrame(data=data)
return df_new_users
def New_Signup_Web_pull(from_date, to_date):
"""
Pull app creator sign up events in Mixpanel.
Parameters
----------
from_date: date
Start date of query.
to_date: date
End date of query.
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe contains user IDs and new sign up event count for app creators in Mixpanel.
"""
#generate JQL query
query = JQL(api_creator_secret,
events=Events({
'event_selectors': [{
'event': "New Signup Web"
}],
'from_date': from_date,
'to_date': to_date
})).group_by(
keys=[
"e.properties.userId", #use userId
],
accumulator=Reducer.count())
#initialize lists to record user IDs and New Signup Web
user_id_list = []
new_sign_up_list = []
for row in query.send():
if row['key'][0] is not None:
user_id_list.append(int(row['key'][0]))
new_sign_up_list.append(row['value'])
#generate dataframe
data = {'user_id': user_id_list, 'new_sign_up': new_sign_up_list}
df_New_Signup_Web = pd.DataFrame(data)
return df_New_Signup_Web
def creator_pull():
"""
Pull app creators' info in Mixpanel.
Parameters
----------
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe app creators info in Mixpanel.
"""
#generate JQL query
query = JQL(api_creator_secret,
people=People({'user_selectors':
[{}]})).group_by(keys=[
"e.properties.$email",
"e.properties.$username",
"e.properties.hs_owner"
],
accumulator=Reducer.count())
#store emails, user IDs, and user journey stages in lists
email_list = []
user_id_list = []
hs_owner_list = []
for row in query.send():
if row['key'][0] is not None:
email_list.append(row['key'][0])
user_id_list.append(int(row['key'][1]))
hs_owner_list.append(row['key'][2])
#create dataframe
data = {
'email': email_list,
'user_id': user_id_list,
'hs_owner': hs_owner_list
}
df_creators = pd.DataFrame(data=data)
#only keep users with HubSpot owner field. Indicating that it exists in hubspot
df_creators = df_creators[~df_creators.hs_owner.isnull()]
return df_creators
def app_creator_pull():
"""
Pull app creators' info from Mixpanel.
Parameters
----------
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe app creators info in Mixpanel.
"""
#generate JQL query
query_user = JQL(api_creator_secret,
people=People({'user_selectors': [{}]
})).group_by(keys=[
"e.properties.$email",
"e.properties.$username",
"e.properties.$distinct_id",
"e.properties.active_milestone"
],
accumulator=Reducer.count())
#store emails, user IDs, distinct_id, active_milestone in lists
email_list = []
user_id_list = []
distinct_id_list = []
active_milestone_list = []
for row in query_user.send():
if row['key'][1] is not None:
email_list.append(row['key'][0])
user_id_list.append(int(row['key'][1]))
distinct_id_list.append(row['key'][2])
active_milestone_list.append(row['key'][3])
#create dataframe
data = {
'email': email_list,
'app_owner_id': user_id_list,
'distinct_id': distinct_id_list,
'active_milestone': active_milestone_list
}
df_creators = pd.DataFrame(data=data)
return df_creators
def user_pull():
"""
Pull app creators' info in Mixpanel.
Parameters
----------
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe contains app creators info in Mixpanel.
"""
#generate JQL query
query = JQL(api_creator_secret,
people=People({'user_selectors':
[{}]})).group_by(keys=[
"e.properties.$email",
"e.properties.company_domain",
"e.properties.$distinct_id"
],
accumulator=Reducer.count())
#initiate list to store emails, user IDs, and company domains
email_list = []
company_domain_list = []
distinct_id_list = []
#process query results
for row in query.send():
email_list.append(row['key'][0])
company_domain_list.append(row['key'][1])
distinct_id_list.append(row['key'][2])
#create dataframe
data = {
'email': email_list,
'company_domain': company_domain_list,
'distinct_id': distinct_id_list
}
df_users = pd.DataFrame(data=data)
return df_users
def creator_pull():
"""
Pull app creators' info in Mixpanel.
Parameters
----------
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe contains app creators info in Mixpanel.
"""
#generate JQL query
query = JQL(api_creator_secret,
people=People({'user_selectors':
[{}]})).group_by(keys=[
"e.properties.$email",
"e.properties.$username",
"e.properties.user_journey_stage"
],
accumulator=Reducer.count())
#initialize lists to store user IDs, email, and journey stage
email_list = []
user_id_list = []
user_journey_stage_list = []
#process query results
for row in query.send():
email_list.append(row['key'][0])
user_id_list.append(row['key'][1])
user_journey_stage_list.append(row['key'][2])
#create dataframe
data = {
'email': email_list,
'creator_user_id': user_id_list,
'user_journey_stage': user_journey_stage_list
}
df_creators = pd.DataFrame(data=data)
return df_creators
def creator_pull():
"""
Pull app creators' info in Mixpanel.
Parameters
----------
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe app creators info in Mixpanel.
"""
#generate JQL query
query_user = JQL(
api_creator_secret,
people=People({
'user_selectors': [{
}
]
})
).group_by(
keys=[
"e.properties.$distinct_id",
"e.properties.$username"],
accumulator=Reducer.count()
)
#store emails, user IDs, and user journey stages in lists
distinct_id_list = []
user_id_list = []
for row in query_user.send():
distinct_id_list.append(row['key'][0])
user_id_list.append(row['key'][1])
#create dataframe
data = {'distinct_id': distinct_id_list, 'user_id': user_id_list}
df_creators = pd.DataFrame(data=data)
return df_creators
def pull_creators():
"""
Pull app creators Mixpanel information.
Parameters
----------
Global Variables
----------
api_creator_secret: Mixpanel Creator Project API key
Use to make API calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe containing app creator information from Mixpanel.
"""
#generate JQL query
query_category = JQL(
api_creator_secret, people=People({'user_selectors': [{}]})).group_by(
keys=["e.properties.$email", "e.properties.$distinct_id"],
accumulator=Reducer.count())
#initialize lists to store emails and distinct IDs
email_list = []
distinct_id_list = []
#process query response
for row in query_category.send():
email_list.append(row['key'][0])
distinct_id_list.append(row['key'][1])
#create dataframe
data = {'email': email_list, 'mixpanel_distinct_id': distinct_id_list}
df_creators_mixpanel = pd.DataFrame(data=data)
#remove creators with missing information
df_creators_mixpanel = df_creators_mixpanel.dropna()
return df_creators_mixpanel
def user_pull():
"""
Pull app creators Mixpanel information.
Parameters
----------
Global Variables
----------
api_creator_secret: Mixpanel Creator Project API key.
Use to make API calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe containing app creator information from Mixpanel.
"""
#generate JQL query
query_user = JQL(api_creator_secret,
people=People({'user_selectors': [{}]
})).group_by(keys=[
"e.properties.$email",
"e.properties.$username"
],
accumulator=Reducer.count())
#initialize lists to record app creator information
email_list = []
user_id_list = []
for row in query_user.send():
email_list.append(row['key'][0])
user_id_list.append(row['key'][1])
#create dataframe
data = {'email': email_list, 'user_id': user_id_list}
df_users = pd.DataFrame(data=data)
return df_users
def creator_pull():
"""
Pull app creators' info in Mixpanel.
Parameters
----------
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe app creators info in Mixpanel.
"""
#generate JQL query
query_user = JQL(api_creator_secret,
people=People({'user_selectors': [{}]
})).group_by(keys=[
"e.properties.$email",
"e.properties.$unsubscribed"
],
accumulator=Reducer.count())
#store emails, user IDs, and user journey stages in lists
email_list = []
unsubscribed_list = []
for row in query_user.send():
if ((row['key'][0] is not None) & (row['key'][1] is not None)
): #only keep accounts with both email and unsubscribe status
email_list.append(row['key'][0])
unsubscribed_list.append(row['key'][1])
#create dataframe
data = {'email': email_list, 'unsubscribed': unsubscribed_list}
df_creators = pd.DataFrame(data=data)
return df_creators
def app_user_pull():
"""
Pull app users' information in Mixpanel.
Parameters
----------
Global Variables
----------
api_user_secret: str
Client secret used to make calls to Mixpanel User Project.
Returns
----------
dataframe
Dataframe contains app users' information in Mixpanel.
"""
#generate JQL query
query_user = JQL(api_user_secret, people=People({
'user_selectors': [{}]
})).group_by(keys=["e.properties.$email", "e.properties.$username"],
accumulator=Reducer.count())
#initialize lists to record email and user ID
email_list = []
user_id_list = []
#process query results
for row in query_user.send():
email_list.append(row['key'][0])
user_id_list.append(row['key'][1])
#create dataframe
data = {'user_email': email_list, 'app_user_id': user_id_list}
df_app_users = pd.DataFrame(data=data)
return df_app_users
def creator_pull():
#generate JQL query
query = JQL(api_creator_secret,
people=People({'user_selectors': [{}]})).group_by(
keys=["e.properties.$email", "e.properties.$distinct_id"],
accumulator=Reducer.count())
#store emails, user IDs, and user journey stages in lists
email_list = []
distinct_id_list = []
for row in query.send():
if row['key'][0] is not None:
email_list.append(row['key'][0])
distinct_id_list.append(row['key'][1])
#create dataframe
data = {'email': email_list, 'distinct_id': distinct_id_list}
df_creators = pd.DataFrame(data=data)
return df_creators
def user_appstart_pull(from_date, to_date):
"""
Pull app user AppStart events in Mixpanel.
Parameters
----------
from_date: date
Start date of query.
to_date: date
End date of query.
Global Variables
----------
api_user_secret: str
Client secret used to make calls to Mixpanel User Project.
Returns
----------
df_user_AppStart: dataframe
Dataframe contains app creator user ID and number of app users.
"""
#generate JQL query
query = JQL(api_user_secret,
events=Events({
'event_selectors': [{
'event': "AppStart"
}],
'from_date': from_date,
'to_date': to_date
})).group_by(
keys=["e.properties.zUserId", "e.properties.zAppOwnerId"],
accumulator=Reducer.count())
#initialize lists to record app user IDs, app creator IDs, app name, and number of AppStarts
app_user_id_list = []
owner_id_list = []
AppStart_list = []
#process query results
for row in query.send():
if (row['key'][0] is not None) & (row['key'][1] is not None):
app_user_id_list.append(int(row['key'][0]))
owner_id_list.append(int(row['key'][1]))
AppStart_list.append(row['value'])
#generate email
data = {
'app_user_id': app_user_id_list,
'creator_id': owner_id_list,
'AppStart': AppStart_list
}
df_AppStart = pd.DataFrame(data)
#make sure IDs are valid
df_user_AppStart = df_AppStart[(df_AppStart.app_user_id > 1)
& (df_AppStart.creator_id > 1)]
return df_user_AppStart
def EditorAction_pull(from_date, to_date):
"""
Pull app creator EditorAction-Save events in Mixpanel.
Parameters
----------
from_date: date
Start date of query.
to_date: date
End date of query.
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe contains app creator emails, number of EditorAction events, and zEvent = Save filter.
"""
#generate JQL query
query = JQL(api_creator_secret,
events=Events({
'event_selectors': [{
'event': "EditorAction"
}],
'from_date': from_date,
'to_date': to_date
})).group_by(
keys=["e.properties.zUserEmail", "e.properties.zEvent"],
accumulator=Reducer.count())
#initialize lists to record emails, EditorAction events, and zEvent
user_email_list = []
EditorAction_list = []
zevent_list = []
#process query results
for row in query.send():
if row['key'][0] is not None:
user_email_list.append(row['key'][0])
zevent_list.append(row['key'][1])
EditorAction_list.append(row['value'])
#create dataframe
data = {
'email': user_email_list,
'zevent': zevent_list,
'EditorAction': EditorAction_list
}
df_editor_action = pd.DataFrame(data)
#filter to only include "Save" editor events
df_editor_action = df_editor_action[df_editor_action.zevent == 'Save']
return df_editor_action
def app_user_pull():
"""
Pull app users' info in Mixpanel.
Parameters
----------
Global Variables
----------
api_user_secret: str
Client secret used to make calls to Mixpanel User Project.
Returns
----------
dataframe
Dataframe contains app users info in Mixpanel.
"""
#generate JQL query
query = JQL(api_user_secret, people=People(
{'user_selectors': [{}]})).group_by(
keys=[
"e.properties.$email",
"e.properties.$username",
"e.properties.$distinct_id",
"e.properties.creator", #pulling current creator status
"e.properties.active_user"
], #pulling current active user status
accumulator=Reducer.count())
#initialize lists to record email, user ID, creator status, active user status, and distinct ID
email_list = []
user_id_list = []
creator_list = []
active_user_list = []
distinct_id_list = []
#process query results
for row in query.send():
email_list.append(row['key'][0])
user_id_list.append(row['key'][1])
distinct_id_list.append(row['key'][2])
creator_list.append(row['key'][3])
active_user_list.append(row['key'][4])
#create dataframe
data = {
'email': email_list,
'app_user_id': user_id_list,
'distinct_id': distinct_id_list,
'creator_current': creator_list,
'active_user_current': active_user_list
}
df_app_users = pd.DataFrame(data=data)
df_app_users = df_app_users.dropna()
df_app_users.app_user_id = df_app_users.app_user_id.astype(int)
return df_app_users
def get_users_stats_from_mixpanel(user_data_dict,
is_single_user=False,
include_stats=False):
"""
This method will fetch user stats from MixPanel using JQL and Candidate Table using SQL
:param user_data_dict: Dict containing data for all users in system
:param is_single_user: Are we getting stats for a single user
:param include_stats: Include statistics of user in response
:return: Dict containing data for all users in system
:rtype: dict
"""
if not include_stats:
return user_data_dict
if is_single_user:
user_data_dict['candidates_count'] = 0
user_data_dict['logins_per_month'] = 0
user_data_dict['searches_per_month'] = 0
user_data_dict['campaigns_count'] = 0
user_data_dict['pipelines_count'] = 0
user_data_dict['emails_count'] = 0
else:
for user_id, user_data in user_data_dict.iteritems():
user_data['candidates_count'] = 0
user_data['logins_per_month'] = 0
user_data['searches_per_month'] = 0
user_data['campaigns_count'] = 0
user_data['pipelines_count'] = 0
user_data['emails_count'] = 0
request_origin = request.environ.get('HTTP_ORIGIN', '')
logger.info('Request Origin for users GET request is: %s', request_origin)
if not request_origin:
url_prefix = 'staging.gettalent' if app.config[
TalentConfigKeys.ENV_KEY] in (
TalentEnvs.QA, TalentEnvs.DEV,
TalentEnvs.JENKINS) else 'app.gettalent'
else:
parsed_url = urlparse(request_origin)
url_prefix = parsed_url.netloc
to_date = datetime.utcnow()
from_date = to_date - timedelta(days=30)
if is_single_user:
selector = '"{}" in properties["$current_url"] and properties["id"] == {}'.format(
url_prefix, user_data_dict['id'])
else:
selector = '"{}" in properties["$current_url"]'.format(url_prefix)
params = {
'event_selectors': [{
'event': 'Login',
'selector': selector
}, {
'event': 'Search',
'selector': selector
}],
'from_date':
str(from_date.date()),
'to_date':
str(to_date.date())
}
try:
query = JQL(app.config[TalentConfigKeys.MIXPANEL_API_KEY],
params).group_by(keys=["e.properties.id", "e.name"],
accumulator=Reducer.count())
iterator = query.send()
except Exception as e:
logger.error(
"Error while fetching user stats from MixPanel because: %s" %
e.message)
raise InvalidUsage("Error while fetching user stats")
for row in iterator:
user_dict_key = 'logins_per_month' if row['key'][
1] == 'Login' else 'searches_per_month'
if is_single_user and row['key'][0] == user_data_dict['id']:
user_data_dict[user_dict_key] = row['value']
elif (not is_single_user) and (row['key'][0] in user_data_dict):
user_data_dict[row['key'][0]][user_dict_key] = row['value']
# Get Candidate, Pipeline and Campaigns Stats of a User
if is_single_user:
user_data_dict['pipelines_count'] = TalentPipeline.query.filter_by(
user_id=user_data_dict['id']).count()
user_data_dict['campaigns_count'] = EmailCampaign.query.filter_by(
user_id=user_data_dict['id']).count()
user_data_dict['candidates_count'] = Candidate.query.filter_by(
user_id=user_data_dict['id']).count()
user_data_dict['emails_count'] = EmailCampaignSend.query.join(
EmailCampaign).filter(
EmailCampaign.user_id == user_data_dict['id']).count()
else:
users_candidate_count = db.session.query(
Candidate.user_id,
func.count(Candidate.user_id)).group_by(Candidate.user_id).all()
users_pipelines_count = db.session.query(
TalentPipeline.user_id,
func.count(TalentPipeline.user_id)).group_by(
TalentPipeline.user_id).all()
users_campaigns_count = db.session.query(
EmailCampaign.user_id, func.count(EmailCampaign.user_id)).group_by(
EmailCampaign.user_id).all()
users_emails_count = EmailCampaignSend.query.join(
EmailCampaign).with_entities(EmailCampaign.user_id,
func.count(
EmailCampaignSend.id)).group_by(
EmailCampaign.user_id).all()
for user_id, count in users_candidate_count:
if user_id in user_data_dict:
user_data_dict[user_id]['candidates_count'] = count
for user_id, count in users_pipelines_count:
if user_id in user_data_dict:
user_data_dict[user_id]['pipelines_count'] = count
for user_id, count in users_campaigns_count:
if user_id in user_data_dict:
user_data_dict[user_id]['campaigns_count'] = count
for user_id, count in users_emails_count:
if user_id in user_data_dict:
user_data_dict[user_id]['emails_count'] = count
return user_data_dict
def New_Signup_Web_pull(from_date, to_date):
"""
Pull app creator sign up events in Mixpanel.
Parameters
----------
from_date: date
Start date of query.
to_date: date
End date of query.
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe contains user IDs, sign up datetime, new sign up event count, and country for app creators in Mixpanel.
"""
#generate JQL query
query = JQL(api_creator_secret,
events=Events({
'event_selectors': [{
'event': "New Signup Web"
}],
'from_date': from_date,
'to_date': to_date
})).group_by(keys=[
"e.properties.userId", "new Date(e.time).toISOString()"
],
accumulator=Reducer.count())
#store emails, user IDs, sign up datetime, and country
user_id_list = []
datetime_list = []
new_sign_up_list = []
for row in query.send():
if row['key'][0] is not None:
user_id_list.append(int(row['key'][0]))
datetime_list.append(
datetime.strptime(row['key'][1][:10], '%Y-%m-%d'))
new_sign_up_list.append(row['value'])
#generate dataframe
data = {
'app_owner_id': user_id_list,
'date': datetime_list,
'new_sign_up': new_sign_up_list
}
df_New_Signup_Web = pd.DataFrame(data)
#only keeping ones with actual IDs
df_New_Signup_Web = df_New_Signup_Web[df_New_Signup_Web.app_owner_id > 1]
return df_New_Signup_Web
def Editor_pull(from_date, to_date):
"""
Pull app creator Editor events in Mixpanel.
Parameters
----------
from_date: date
Start date of query.
to_date: date
End date of query.
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe contains user IDs, Editor event datetime, and Editor event count for app creators in Mixpanel.
"""
#generate JQL query
query = JQL(api_creator_secret,
events=Events({
'event_selectors': [{
'event': "Editor"
}],
'from_date': from_date,
'to_date': to_date
})).group_by(keys=[
"e.properties.zUserId", "new Date(e.time).toISOString()"
],
accumulator=Reducer.count())
#store user IDs, Editor count, and Editor datetime
user_id_list = []
datetime_list = []
Editor_list = []
for row in query.send():
if row['key'][0] is not None:
user_id_list.append(int(row['key'][0]))
datetime_list.append(
datetime.strptime(row['key'][1][:10], '%Y-%m-%d'))
Editor_list.append(row['value'])
#generate dataframe
data = {
'user_id': user_id_list,
'Editor_datetime': datetime_list,
'Editor': Editor_list
}
df_Editor = pd.DataFrame(data)
return df_Editor
def user_appstart_pull(from_date, to_date):
"""
Pull app user AppStart events in Mixpanel.
Parameters
----------
from_date: date
Start date of query.
to_date: date
End date of query.
Global Variables
----------
api_user_secret: str
Client secret used to make calls to Mixpanel User Project.
Returns
----------
df_user_AppStart: dataframe
Dataframe contains app creator user ID and number of app users.
"""
#generate JQL query
query = JQL(api_user_secret,
events=Events({
'event_selectors': [{
'event': "AppStart"
}],
'from_date': from_date,
'to_date': to_date
})).group_by(keys=[
"e.properties.zUserId", "e.properties.zAppOwnerId",
"new Date(e.time).toISOString()"
],
accumulator=Reducer.count())
#initialize lists to record app user IDs, app creator IDs, app name, and number of AppStarts
app_user_id_list = []
owner_id_list = []
AppStart_list = []
date_list = []
#process query results
for row in query.send():
if (row['key'][0] is not None) & (row['key'][1] is not None):
app_user_id_list.append(int(row['key'][0]))
owner_id_list.append(int(row['key'][1]))
date_list.append(datetime.strptime(row['key'][2][:10], '%Y-%m-%d'))
AppStart_list.append(row['value'])
#generate email
data = {
'date': date_list,
'app_user_id': app_user_id_list,
'user_id': owner_id_list,
'AppStart': AppStart_list
}
df_AppStart = pd.DataFrame(data)
#make sure IDs are valid
df_AppStart = df_AppStart[(df_AppStart.app_user_id > 1)
& (df_AppStart.user_id > 1)]
#remove duplicate users on the same day. Since it was pulled based on timestamp, AppStarts in one day can be on multiple roads
df_AppStart = df_AppStart.drop_duplicates(
['user_id', 'date', 'app_user_id'])
#get total users and list of user emails for each app creator
df_user_AppStart = df_AppStart.groupby(
['user_id', 'date']).app_user_id.count().reset_index()
df_user_AppStart = df_user_AppStart.rename(
columns={'app_user_id': 'num_app_users'})
return df_user_AppStart
def New_Signup_App_pull(yesterday):
"""
Pull app user sign up events in Mixpanel.
Parameters
----------
yesterday: date
Yesterday's date.
Global Variables
----------
api_user_secret: str
Client secret used to make calls to Mixpanel User Project.
Returns
----------
dataframe
Dataframe contains user IDs, sign up datetime, and new sign up event count for app users in Mixpanel.
"""
#generate JQL query
query = JQL(api_user_secret,
events=Events({
'event_selectors': [{
'event': "New Signup App"
}],
'from_date': yesterday,
'to_date': yesterday
})).group_by(keys=[
"e.properties.zUserId", "new Date(e.time).toISOString()"
],
accumulator=Reducer.count())
#initialize lists to store emails, user IDs, and sign up datetime
user_id_list = []
datetime_list = []
new_sign_up_list = []
#process query results
for row in query.send():
if row['key'][0] is not None:
user_id_list.append(int(row['key'][0]))
datetime_list.append(
datetime.strptime(row['key'][1][:10], '%Y-%m-%d'))
new_sign_up_list.append(row['value'])
#generate dataframe
data = {
'app_user_id': user_id_list,
'sign_up_datetime': datetime_list,
'new_signup_app': new_sign_up_list
}
df_New_Signup_App = pd.DataFrame(data)
return df_New_Signup_App
def pull_usage(from_date, to_date):
"""
Pull usage events in Mixpanel.
Parameters
----------
from_date: date
Start date of query.
to_date: date
End date of query.
Global Variables
----------
api_creator_secret: str
Client secret used to make calls to Mixpanel Creator Project.
Returns
----------
dataframe
Dataframe contains app ID, user IDs, owner ID, and usage event count in Mixpanel.
"""
#generate JQL query
query = JQL(api_creator_secret,
events=Events({
'event_selectors': [{
'event': "Usage"
}],
'from_date': from_date,
'to_date': to_date
})).group_by(
keys=["e.properties.OwnerId", "e.properties.UserId"],
accumulator=Reducer.count())
#store app owner, app, and app user IDs
app_owner_id_list = []
app_user_id_list = []
usage_list = []
for row in query.send():
if row['key'][0] is not None:
app_owner_id_list.append(int(row['key'][0]))
app_user_id_list.append(int(row['key'][1]))
usage_list.append(row['value'])
#generate dataframe
data = {
'app_owner_id': app_owner_id_list,
'app_user_id': app_user_id_list,
'usage': usage_list
}
df_usage = pd.DataFrame(data)
#only keep app owners and users with proper IDs
df_usage = df_usage[(df_usage.app_owner_id > 1)
& (df_usage.app_user_id > 1)]
df_usage = df_usage[~df_usage.app_owner_id.isin(
[10305, 71626])] #remove for demo accounts
return df_usage
def user_appstart_pull(from_date, to_date, df_app_users):
"""
Pull app user AppStart events in Mixpanel.
Parameters
----------
from_date: date
Start date of query.
to_date: date
End date of query.
Global Variables
----------
api_user_secret: str
Client secret used to make calls to Mixpanel User Project.
Returns
----------
df_user_AppStart: dataframe
Dataframe contains app creator user ID, list of app user email domains, and number of app users
df_top_appstarts: dataframe
Dataframe contains app creator user ID and number of AppStarts from top apps
"""
#generate JQL query
query = JQL(api_user_secret,
events=Events({
'event_selectors': [{
'event': "AppStart"
}],
'from_date': from_date,
'to_date': to_date
})).group_by(keys=[
"e.properties.zUserId", "e.properties.zAppOwnerId",
"e.properties.zAppName"
],
accumulator=Reducer.count())
#initialize lists to record app user IDs, app creator IDs, app name, and number of AppStarts
app_user_id_list = []
owner_id_list = []
app_name_list = []
AppStart_list = []
#process query results
for row in query.send():
if row['key'][0] is not None:
app_user_id_list.append(row['key'][0])
owner_id_list.append(row['key'][1])
app_name_list.append(row['key'][2])
AppStart_list.append(row['value'])
#generate email
data = {
'app_user_id': app_user_id_list,
'user_id': owner_id_list,
'app_name': app_name_list,
'AppStart': AppStart_list
}
df_AppStart = pd.DataFrame(data)
#merge AppStart and app user dataframes to associate app user emails to number of AppStarts
df_AppStart = pd.merge(df_AppStart,
df_app_users,
on='app_user_id',
how='left')
#get total users and list of user emails for each app creator
df_user_AppStart = df_AppStart.groupby(
['user_id', 'app_user_id'])['app_name'].count().reset_index()
df_user_AppStart = df_user_AppStart.groupby(
'user_id')['app_user_id'].count().reset_index()
df_user_AppStart = df_user_AppStart.rename(
columns={'app_user_id': 'num_app_users'})
#get app user domain list for each app creator
df_AppStart_temp = df_AppStart.groupby(
['user_id', 'app_user_id',
'user_email'])['app_name'].count().reset_index()
df_AppStart_temp['user_email_domains'] = df_AppStart_temp[
'user_email'].str.split('@').str[1].fillna('')
df_AppStart_temp = df_AppStart_temp[
df_AppStart_temp['user_email_domains'] != '']
df_user_domains_by_creators = df_AppStart_temp.groupby(
'user_id')['user_email_domains'].apply(list).reset_index()
df_user_AppStart = pd.merge(df_user_AppStart,
df_user_domains_by_creators,
on='user_id',
how='left')
df_user_AppStart.loc[df_user_AppStart['user_email_domains'].isnull(),
['user_email_domains']] = df_user_AppStart.loc[
df_user_AppStart['user_email_domains'].isnull(),
'user_email_domains'].apply(lambda x: [])
#only keep the app with the highest number of AppStarts for each app creator
df_top_appstarts = df_AppStart.groupby(
['user_id', 'app_name']).AppStart.sum().reset_index().sort_values(
'AppStart', ascending=False).drop_duplicates('user_id',
keep='first')
df_top_appstarts = df_top_appstarts.rename(
columns={'AppStart': 'appstart_by_top_app'})
return df_user_AppStart, df_top_appstarts