b u ilding f u nctions to a u tomate anal y sis
play

B u ilding f u nctions to a u tomate anal y sis AN ALYZIN G MAR K E - PowerPoint PPT Presentation

Mar 24, 2023 •36 likes •306 views

B u ilding f u nctions to a u tomate anal y sis AN ALYZIN G MAR K E TIN G C AMPAIG N S W ITH PAN DAS Jill Rosok Data Scientist Wh y b u ild a f u nction ? # Count the subs by referring channel and day retention_total =

  1. B u ilding f u nctions to a u tomate anal y sis AN ALYZIN G MAR K E TIN G C AMPAIG N S W ITH PAN DAS Jill Rosok Data Scientist

  2. Wh y b u ild a f u nction ? # Count the subs by referring channel and day retention_total = marketing.groupby(['date_subscribed', 'subscribing_channel'])\ ['user_id'].nunique() retention_subs = marketing[marketing['is_retained'] == True]\ .groupby(['date_subscribed', 'subscribing_channel'])\ ['user_id'].nunique() # Calculate the retention rate daily_retention_rate = retention_subs/retention_total daily_retention_rate = pd.DataFrame( daily_retention_rate.unstack(level=1) ) ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  3. print(daily_retention_rate) subscribing_channel Email Facebook House Ads Instagram Push date_subscribed 2018-01-01 1.00 0.875000 0.687500 0.750000 1.000000 2018-01-02 0.75 1.000000 0.588235 0.625000 1.000000 2018-01-03 NaN 0.800000 0.647059 0.909091 0.666667 2018-01-04 1.00 0.666667 0.466667 0.500000 NaN 2018-01-05 1.00 0.571429 0.500000 0.636364 1.000000 ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  4. B u ilding a retention f u nction def retention_rate(dataframe, column_names): # Group by column_names and calculate retention retained = dataframe[dataframe['is_retained'] == True]\ .groupby(column_names)['user_id'].nunique() # Group by column_names and calculate conversion converted = dataframe[dataframe['converted'] == True]\ .groupby(column_names)['user_id'].nunique() retention_rate = retained/converted return retention_rate ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  5. Retention rate b y channel daily_retention = retention_rate(marketing, ['date_subscribed', 'subscribing_channel']) daily_retention = pd.DataFrame( daily_retention.unstack(level=1) ) print(daily_retention.head()) subscribing_channel Email Facebook House Ads Instagram Push date_subscribed 2018-01-01 1.00 0.875000 0.687500 0.750000 1.000000 2018-01-02 0.75 1.000000 0.588235 0.625000 1.000000 2018-01-03 NaN 0.800000 0.647059 0.909091 0.666667 2018-01-04 1.00 0.666667 0.466667 0.500000 NaN 2018-01-05 1.00 0.571429 0.500000 0.636364 1.000000 ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  6. Plotting dail y retention b y channel daily_retention.plot(date_subscribed, conversion_rate) plt.title('Daily channel retention rate\n', size = 16) plt.ylabel('Retention rate (%)', size = 14) plt.xlabel('Date', size = 14) plt.show() ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  7. ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  8. Plotting f u nction def plotting(dataframe): for column in dataframe: plt.plot(dataframe.index, dataframe[column]) plt.title('Daily ' + column + ' retention rate\n', size = 16) plt.ylabel('Retention rate (%)', size = 14) plt.xlabel('Date', size = 14) plt.show() plotting(daily_channel_retention) ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  9. ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  10. Let ' s practice ! AN ALYZIN G MAR K E TIN G C AMPAIG N S W ITH PAN DAS

  11. Identif y ing inconsistencies AN ALYZIN G MAR K E TIN G C AMPAIG N S W ITH PAN DAS Jill Rosok Data Scientist

  12. Da y of w eek trends DoW_retention = retention_rate(marketing, ['DoW']) ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  13. Plotting the res u lts # Plot retention by day of week DoW_retention.plot() plt.title('Retention rate by day of week') plt.ylim(0) plt.show() ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  14. ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  15. Real data can be mess y and conf u sing ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  16. Let ' s practice ! AN ALYZIN G MAR K E TIN G C AMPAIG N S W ITH PAN DAS

  17. Resol v ing inconsistencies AN ALYZIN G MAR K E TIN G C AMPAIG N S W ITH PAN DAS Jill Rosok Data Scientist

  18. Assessing impact # Calculate pre-error conversion rate # Bug arose sometime around '2018-01-11' house_ads_no_bug = house_ads[house_ads['date_served'] < '2018-01-11 lang_conv = conversion_rate(house_ads_no_bug, ['language_displayed']) ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  19. Assessing impact # Index other language conversion rate against English spanish_index = lang_conv['Spanish']/lang_conv['English'] arabic_index = lang_conv['Arabic']/lang_conv['English'] german_index = lang_conv['German']/lang_conv['English'] ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  20. Interpreting Inde x es print("Spanish index:", spanish_index) print("Arabic index:", arabic_index) print("German index:", german_index) Spanish index: 1.6819248826291078 Arabic index: 5.045774647887324 German index: 4.485133020344288 ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  21. Dail y con v ersion # Create actual conversion DataFrame language_conversion = house_ads.groupby(['date_served', \ 'language_preferred'])\ .agg({'user_id':'nunique',\ 'converted':'sum'}) ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  22. Dail y con v ersion expected_conversion = pd.DataFrame( language_conversion.unstack(level=1) ) user_id converted language_preferred Arabic English German Spanish Arabic English German Spanish date_served 2018-01-01 2.0 171.0 5.0 11.0 2 13 1 0 2018-01-02 3.0 200.0 5.0 10.0 0 14 3 0 2018-01-03 2.0 179.0 3.0 8.0 0 15 1 1 2018-01-04 2.0 149.0 2.0 14.0 0 12 0 3 2018-01-05 NaN 143.0 1.0 14.0 NaN 17 False 3 ... ... ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  23. Create English con v ersion rate col u mn # Create English conversion rate column for affected period language_conversion['actual_english_conversions'] = \ language_conversion.loc\ ['2018-01-11':'2018-01-31']\ [('converted','English')] ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  24. Calc u lating dail y e x pected con v ersion rate # Create expected conversion rates for each language language_conversion['expected_spanish_rate'] = \ language_conversion['actual_english_rate']*spanish_index language_conversion['expected_arabic_rate'] = \ language_conversion['actual_english_rate']*arabic_index language_conversion['expected_german_rate'] = \ language_conversion['actual_english_rate']*german_index ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  25. Calc u lating dail y e x pected con v ersions # Multiply total ads served by expected conversion rate language_conversion['expected_spanish_conversions'] = \ language_conversion['expected_spanish_rate']/100 *language_conversion[('user_id','Spanish')] language_conversion['expected_arabic_conversions'] = \ language_conversion['expected_arabic_rate']/100 *language_conversion[('user_id','Arabic')] language_conversion['expected_german_conversions'] = \ language_conversion['expected_german_rate']/100 *language_conversion[('user_id','German')] ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  26. Determining the n u mber of lost s u bscribers bug_period = language_conversion.loc['2018-01-11':'2018-01-31'] # Sum expected subscribers for each language expected_subs = bug_period['expected_spanish_conv_rate'].agg('sum') + \ bug_period['expected_arabic_conv_rate'].agg('sum') + \ bug_period['expected_german_conv_rate'].agg('sum') # Calculate how many subscribers we actually got actual_subs = bug_period[('converted','Spanish')].sum() + \ bug_period[('converted','Arabic')].agg('sum') + \ bug_period[('converted','German')].agg('sum') lost_subs = expected_subs - actual_subs print(lost_subs) 32.144143192488265 ANALYZING MARKETING CAMPAIGNS WITH PANDAS

  27. Let ' s practice ! AN ALYZIN G MAR K E TIN G C AMPAIG N S W ITH PAN DAS

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

Anal nalys ysis is of Bu Bus s Coll llision ision In Incident idents s Us Usin ing g

Anal nalys ysis is of Bu Bus s Coll llision ision In Incident idents s Us Usin ing g

Anal nalys ysis is of Bu Bus s Coll llision ision In Incident idents s Us Usin ing g Da Data a Dr Driv iven en Tools ols in in De Devel elopi oping ng and and Bu Buil ilding ing a S a Saf afety ty Fi First st Cul

388 views • 15 slides
F u nctions as objects W R ITIN G FU N C TION S IN P YTH ON Sha y ne Miel Director of So w

F u nctions as objects W R ITIN G FU N C TION S IN P YTH ON Sha y ne Miel Director of So w

F u nctions as objects W R ITIN G FU N C TION S IN P YTH ON Sha y ne Miel Director of So w are Engineering @ American E cient F u nctions are j u st another t y pe of object P y thon objects : def x(): pass x = [1, 2, 3] x = {'foo': 42}

686 views • 46 slides
Panel f u nctions DATA VISU AL IZATION W ITH L ATTIC E IN R Deepa y an Sarkar Associate

Panel f u nctions DATA VISU AL IZATION W ITH L ATTIC E IN R Deepa y an Sarkar Associate

Panel f u nctions DATA VISU AL IZATION W ITH L ATTIC E IN R Deepa y an Sarkar Associate Professor , Indian Statistical Instit u te Introd u ction E x amples so far u se b u ilt - in high - le v el f u nctions s u ch as histogram() , xyplot() ,

602 views • 30 slides
Lambda f u nctions P YTH ON DATA SC IE N C E TOOL BOX ( PAR T 1 ) H u go Bo w ne - Anderson Instr

Lambda f u nctions P YTH ON DATA SC IE N C E TOOL BOX ( PAR T 1 ) H u go Bo w ne - Anderson Instr

Lambda f u nctions P YTH ON DATA SC IE N C E TOOL BOX ( PAR T 1 ) H u go Bo w ne - Anderson Instr u ctor Lambda f u nctions raise_to_power = lambda x, y: x ** y raise_to_power(2, 3) 8 PYTHON DATA SCIENCE TOOLBOX ( PART 1) Anon y mo u s f u

547 views • 23 slides
F u ndamental Val u ation : Anal yz ing Projections E QU ITY VAL U ATION IN R Cli Ang

F u ndamental Val u ation : Anal yz ing Projections E QU ITY VAL U ATION IN R Cli Ang

F u ndamental Val u ation : Anal yz ing Projections E QU ITY VAL U ATION IN R Cli Ang Senior Vice President , Compass Le x econ Anal yz ing Projections Projections are critical inp u t in disco u nted cash o w anal y sis Garbage - In

261 views • 14 slides
Probabilit y densit y f u nctions STATISTIC AL TH IN K IN G IN P YTH ON ( PAR T 1 ) J u stin

Probabilit y densit y f u nctions STATISTIC AL TH IN K IN G IN P YTH ON ( PAR T 1 ) J u stin

Probabilit y densit y f u nctions STATISTIC AL TH IN K IN G IN P YTH ON ( PAR T 1 ) J u stin Bois Lect u rer at the California Instit u te of Technolog y Contin u o u s v ariables Q u antities that can take an y v al u e , not j u st discrete v

1.17k views • 45 slides
Cen Central Florida Florida Auto tomate mated Ve Vehicle Pr Prov oving Gr Grounds ounds Pa

Cen Central Florida Florida Auto tomate mated Ve Vehicle Pr Prov oving Gr Grounds ounds Pa

Cen Central Florida Florida Auto tomate mated Ve Vehicle Pr Prov oving Gr Grounds ounds Pa Partners rship Overview by Charles A. Ramdatt, P.E., PTOE, AICP for The MetroPlan Orlando Community Advisory Committee February March 2017 Wh

303 views • 14 slides
Creating transcription helper f u nctions SP OK E N L AN G U AG E P R OC E SSIN G IN P YTH ON

Creating transcription helper f u nctions SP OK E N L AN G U AG E P R OC E SSIN G IN P YTH ON

Creating transcription helper f u nctions SP OK E N L AN G U AG E P R OC E SSIN G IN P YTH ON Daniel Bo u rke Machine Learning Engineer / Yo u T u be Creator E x ploring a u dio files # Import os module import os # Check the folder of

816 views • 46 slides
Probabilit y mass f u nctions E XP L OR ATOR Y DATA AN ALYSIS IN P YTH ON Allen Do w ne y

Probabilit y mass f u nctions E XP L OR ATOR Y DATA AN ALYSIS IN P YTH ON Allen Do w ne y

Probabilit y mass f u nctions E XP L OR ATOR Y DATA AN ALYSIS IN P YTH ON Allen Do w ne y Professor , Olin College GSS Ann u al sample of U . S . pop u lation . Asks abo u t demographics , social and political beliefs . Widel y u sed b y polic y

715 views • 37 slides
B u ilding tf - idf doc u ment v ectors FE ATU R E E N G IN E E R IN G FOR N L P IN P YTH ON

B u ilding tf - idf doc u ment v ectors FE ATU R E E N G IN E E R IN G FOR N L P IN P YTH ON

B u ilding tf - idf doc u ment v ectors FE ATU R E E N G IN E E R IN G FOR N L P IN P YTH ON Ro u nak Banik Data Scientist n - gram modeling Weight of dimension dependent on the freq u enc y of the w ord corresponding to the dimension . Doc u

525 views • 41 slides
B u ilding a bag of w ords model FE ATU R E E N G IN E E R IN G FOR N L P IN P YTH ON Ro u

B u ilding a bag of w ords model FE ATU R E E N G IN E E R IN G FOR N L P IN P YTH ON Ro u

B u ilding a bag of w ords model FE ATU R E E N G IN E E R IN G FOR N L P IN P YTH ON Ro u nak Banik Data Scientist Recap of data format for ML algorithms For an y ML algorithm , Data m u st be in tab u lar form Training feat u res m u st be

426 views • 25 slides
Building ilding an an op open en con oncordancer ordancer for or Mal alay ay/In

Building ilding an an op open en con oncordancer ordancer for or Mal alay ay/In

Building ilding an an op open en con oncordancer ordancer for or Mal alay ay/In /Indonesian donesian Shiro Akasegawa Asako Shiohara Hiroki Nomoto Tokyo University of Foreign Studies Lago Institute of Language ISMIL

428 views • 32 slides
H OW I MPORTANT IS THE E XPANSION OF THE P ANAMA C ANAL TO THE P ORT OF N EW O RLEANS Bruce

H OW I MPORTANT IS THE E XPANSION OF THE P ANAMA C ANAL TO THE P ORT OF N EW O RLEANS Bruce

H OW I MPORTANT IS THE E XPANSION OF THE P ANAMA C ANAL TO THE P ORT OF N EW O RLEANS Bruce Lambert Executive Director, ITTS W ARNING P ANAMA C ANAL E XPANSION Is it needed? Does Louisiana Benefit from the Canal now? Does it

400 views • 36 slides
Shado w price sensiti v it y anal y sis SU P P LY C H AIN AN ALYTIC S IN P YTH ON Aaren St u

Shado w price sensiti v it y anal y sis SU P P LY C H AIN AN ALYTIC S IN P YTH ON Aaren St u

Shado w price sensiti v it y anal y sis SU P P LY C H AIN AN ALYTIC S IN P YTH ON Aaren St u bber eld S u ppl y Chain Anal y tics Mgr . Define shado w price Modeling in iss u es : Inp u t for model constraints are o en estimates Will

744 views • 51 slides
Introd u ction to E x plorator y Data Anal y sis STATISTIC AL TH IN K IN G IN P YTH ON ( PAR T 1

Introd u ction to E x plorator y Data Anal y sis STATISTIC AL TH IN K IN G IN P YTH ON ( PAR T 1

Introd u ction to E x plorator y Data Anal y sis STATISTIC AL TH IN K IN G IN P YTH ON ( PAR T 1 ) J u stin Bois Lect u rer at the California Instit u te of Technolog y E x plorator y data anal y sis The process of organi z ing , plo ing ,

734 views • 45 slides
Fu Func nctions tions on on th the Lat attic tice Huey-Wen Lin University of Washington

Fu Func nctions tions on on th the Lat attic tice Huey-Wen Lin University of Washington

Pa Part rton on Di Dist stribution ribution Fu Func nctions tions on on th the Lat attic tice Huey-Wen Lin University of Washington Huey-Wen Lin Lattice Workshop @ NTU 1 Part rton on Dis istri ributio tion n Functio

565 views • 31 slides
L1023.we b Material Retainer stainless (304), rollers Technical notes stainless (440C). Body

L1023.we b Material Retainer stainless (304), rollers Technical notes stainless (440C). Body

Stainless Cross Roller Slides Linear Tables flanged, nickel plated L1023.we b Material Retainer stainless (304), rollers Technical notes stainless (440C). Body stainless steel (440C), nickel Carriage top parallelism 3 . plated apart

315 views • 3 slides
Creating enduring value through sustainability Our 2030 Real Estate Sustainability Strategy

Creating enduring value through sustainability Our 2030 Real Estate Sustainability Strategy

Creating enduring value through sustainability Our 2030 Real Estate Sustainability Strategy AMP Capital Real Estate Sustainability team Melanie Louie Lina Rico Binowee Chris Nunn Gomez Bayles Head of Reporting &amp; Data

362 views • 15 slides
First Quarter 2020 Financial Results 22 April 2020 Outline 1Q 2020 Key Highlights 3

First Quarter 2020 Financial Results 22 April 2020 Outline 1Q 2020 Key Highlights 3

First Quarter 2020 Financial Results 22 April 2020 Outline 1Q 2020 Key Highlights 3 Navigating the COVID-19 Situation 4 Financial Results 7 Portfolio Review 12 Market Update 17 Additional Information 22 IMPORTANT

655 views • 27 slides
Before, During and After Separating from an Abusive Partner November 24, 2020 Webinar prepared

Before, During and After Separating from an Abusive Partner November 24, 2020 Webinar prepared

Safety Planning Before, During and After Separating from an Abusive Partner November 24, 2020 Webinar prepared for CLEO Disclaimer This webinar is for general information purposes only and is not legal advice. It is not intended to be used

768 views • 48 slides
Address selection in multihomed environments multi6 wg meeting - IETF 61 C. Huitema, R. Draves

Address selection in multihomed environments multi6 wg meeting - IETF 61 C. Huitema, R. Draves

Address selection in multihomed environments multi6 wg meeting - IETF 61 C. Huitema, R. Draves M. Bagnulo Scenario /-- ( A ) ---( ) (site X) ( IPv6 ) ---(C)---(site Y) X \-- ( B ) ---( ) Y A:X1

448 views • 5 slides
Welcome to your home church! A Preview of Coming Attractions Revelation 11:1-14 I . THE

Welcome to your home church! A Preview of Coming Attractions Revelation 11:1-14 I . THE

Welcome to your home church! A Preview of Coming Attractions Revelation 11:1-14 I . THE MEASURMENT OF THE TEMPLE (11:1-2) A. There were 2 temples in Jerusalems past. 1. Solomons temple 2. Herods temple B. There will be one more in

86 views • 7 slides
POLI 203 The End of the Death Penalty Prof. Baumgartner ANONYMOUS SURVEY November 24, 2014

POLI 203 The End of the Death Penalty Prof. Baumgartner ANONYMOUS SURVEY November 24, 2014

POLI 203 The End of the Death Penalty Prof. Baumgartner ANONYMOUS SURVEY November 24, 2014 Results, December 1, 2014 Without giving any identifying information, please respond to the same question you answered at the beginning of the semester,

121 views • 11 slides
Judges 21:25 A 350 Year Cycle Relapse A 350 Year Cycle Relapse Retribution A 350

Judges 21:25 A 350 Year Cycle Relapse A 350 Year Cycle Relapse Retribution A 350

Judges 21:25 A 350 Year Cycle Relapse A 350 Year Cycle Relapse Retribution A 350 Year Cycle Relapse Retribution Repentance A 350 Year Cycle Relapse Rescue Retribution Repentance A 350 Year Cycle

382 views • 12 slides

More recommend