KHAN ACADEMY AP COMPSCI
PRINCIPLES DATA ANALYSIS
QUESTIONS WITH ACTUAL SOLUTIONS
Part dd1:
In dd2016, ddresearchers ddstudied ddthe dddifferences ddbetween ddtwo dddialects ddof
ddEnglish ddused ddin ddtweets ddon ddTwitter.
They ddfirst ddcategorized ddover ddeight ddmillion ddtweets ddas ddeither ddAfrican-American-
aligned dd(AA-aligned) ddor ddWhite-aligned. dd[How?]
The ddresearchers ddthen ddtried ddout ddthree dddifferent ddlanguage ddidentification
ddalgorithms ddto ddsee ddwhether ddeach ddtweet ddwould ddbe ddcategorized ddas ddthe
ddEnglish ddlanguage, dda dddifferent ddlanguage, ddor ddan ddunknown ddlanguage.
The ddfollowing ddtable ddshows ddthe ddproportion ddof ddtweets ddfrom ddeach dddialect
ddthat ddwere ddclassified ddas ddnon-English ddby ddeach ddalgorithm:
AA-alignedWhite-alignedAlgorithm dd113.2%7.6%Algorithm dd28.4%5.9%Algorithm
dd324.4%17.6%
Which ddof ddthe ddfollowing ddstatements ddis ddsupported ddby ddthat dddata? dd- ddAnswer
dd1.All ddthree ddalgorithms ddcontain ddbiases ddthat ddcause ddthem ddto ddmisidentify ddan
ddAA-aligned ddtweet ddas ddnon-English ddmore ddoften ddthan ddthey ddmisidentify ddthe
ddlanguage ddof dda ddWhite-aligned ddtweet.
2.The ddanalytics ddteam ddof dda ddreviews ddwebsite dduses ddthe ddlanguage
ddidentification ddalgorithm ddto dddisplay dda ddpie ddchart ddof ddthe ddnumber ddof ddreviews
ddwritten ddper ddlanguage ddin ddthe ddlast ddyear.
StackOverflow ddis dda ddpopular ddquestion dd& ddanswers ddsite. ddEach ddtime dda dduser
ddasks dda ddnew ddquestion, ddthey ddinsert dda ddrow ddin dda dddatabase ddtable.
Each ddrow ddcontains:
The dduser ddID
The dduser dddisplay ddname
The ddtimestamp ddof ddthe ddquestion
The ddtext ddof ddthe ddquestion
The ddspam ddscore ddof ddthe ddquestion dd(0-5)
Here ddare dda ddfew ddrows ddfrom ddthe ddtable:
user_iddisplay_nametimestampquestionspam_score62038TheAskerator11/27/2012
dd06:15:28How dddo ddI ddgeocode dda ddlat/lng?020394NewCoder12303/12/2015
dd10:55:10Where ddcan ddI ddhost ddmy ddwebsite ddfor ddfree?
136917QuestionErrthing05/04/2014 dd11:34:25Wanna dddownload ddthis ddfree ddfile?3
The ddteam ddwants ddto dddisplay ddquestion ddstatistics ddon ddan ddinternal dddashboard.
, Which ddstatistic ddcan ddnot ddbe ddcalculated ddfrom ddthe ddtable ddof ddquestions? dd-
ddAnswer ddThe dduser ddID ddwith ddthe ddmost ddnumber ddof ddunanswered ddquestions
A ddteam ddof ddscientists ddand ddengineers ddis ddputting ddtogether dda ddresearch ddproject
ddto ddstudy ddwhale ddsounds. ddIn ddorder ddto dddevelop ddthe ddinfrastructure ddfor ddthe
ddproject, ddthey ddneed ddto ddfirst dddetermine ddhow ddmuch dddata ddstorage ddspace
ddtheir ddobservational dddata ddwill ddrequire.
This ddis ddan ddexample ddof dda ddsingle ddobservation:
SoundLocationDate/timeSpecies3 ddminute ddlong ddaudio ddfile63.776871, dd-
171.742193May dd27, dd2019, dd2:23:13 ddPMBeluga
The ddteam ddhopes ddto ddcapture ddthousands ddof ddwhale ddsounds ddfrom ddall ddthe
ddworld's ddoceans.
Which ddpiece ddof dddata ddwill ddincrease ddtheir dddata ddstorage ddneeds ddthe ddmost? dd-
ddAnswer ddRecording ddof ddwhale ddsound
Safiya ddis dda ddsoftware ddengineer ddat dda ddcompany ddthat's dddeveloping ddsoftware
ddfor ddself-driving ddcars. ddShe's ddworking ddon ddsoftware ddthat dduses ddcomputer
ddvision ddand ddmachine ddlearning ddalgorithms ddto dddetect ddpedestrians ddwalking
ddnear ddthe ddcar ddand ddtrigger ddthe ddbrakes ddwhen ddneeded.
After ddtraining ddthe ddalgorithm ddon dda ddlarge dddataset ddof ddtraining dddata dd(videos
ddof ddpedestrians ddwalking ddnear ddcars), ddthey ddtry ddit ddout ddin ddcars ddwith ddbackup
dddrivers. ddThe dddrivers ddreport ddthat ddit dddetected ddmost ddpedestrians, ddbut ddfailed
ddto dddetect ddpeople ddusing ddwheelchairs ddand ddparents ddpushing ddstrollers. ddThe
dddrivers ddhad ddto ddmanually ddrespond ddin ddthose ddcases.
What's ddthe ddbest ddway ddthat ddSafiya ddcan ddimprove ddthe ddmachine ddlearning
ddalgorithm's ddability ddto dddetect ddall ddpedestrians? dd- ddAnswer ddShe ddcan ddadd
ddvideos ddof ddpeople ddusing ddwheelchairs ddand ddstrollers ddinto ddthe ddtraining dddata
ddset dd(perhaps ddcrowd-sourcing ddthem ddif ddthere ddare ddnone ddalready ddavailable).
An ddonline ddcurriculum ddprovider ddoffers ddtheir ddproduct ddto ddtwo ddaudiences:
ddindependent ddlearners dd(self-directed) ddand ddclassroom ddlearners dd(led ddby ddtheir
ddteachers).
They ddwant ddto ddunderstand ddthe dddifferences ddbetween ddthe ddaudiences ddand ddhow
ddthey dduse ddthe ddproduct, ddso ddthey ddsent ddsurveys ddand ddcollected dddata. ddUsers
ddrated ddtheir ddsatisfaction ddwith ddthe ddproduct ddfrom dd1-10, ddwhere dd1 ddis ddleast
ddsatisfied ddand dd10 ddis ddthe ddmost ddsatisfied.
This ddscatter ddplot ddcompares ddthe ddhours ddper ddweek ddspent ddby dda dduser ddto
ddtheir ddrating ddof ddthe ddproduct:
\small{1}1\small{2}2\small{3}3\small{1}1\small{2}2\small{3}3\small{4}4\small{5}5\
small{6}6\small{7}7\small{8}8\small{9}9\small{10}10\text{rating}rating\text{hours/
week}hours/week\purpleE{\text{independent}}independent\greenE{\
text{classroom}}classroom
A ddscatter ddplot ddwith ddhours ddper ddweek ddon ddthe ddx ddaxis ddand ddrating ddon ddthe
ddy ddaxis. ddThe ddx ddaxis ddgoes ddfrom dd0 ddto dd4 ddand ddthe ddy ddaxis ddgoes ddfrom dd0
ddto dd10. ddBoth ddpurple ddand ddgreen dddots ddare ddscattered ddon dd- ddAnswer
PRINCIPLES DATA ANALYSIS
QUESTIONS WITH ACTUAL SOLUTIONS
Part dd1:
In dd2016, ddresearchers ddstudied ddthe dddifferences ddbetween ddtwo dddialects ddof
ddEnglish ddused ddin ddtweets ddon ddTwitter.
They ddfirst ddcategorized ddover ddeight ddmillion ddtweets ddas ddeither ddAfrican-American-
aligned dd(AA-aligned) ddor ddWhite-aligned. dd[How?]
The ddresearchers ddthen ddtried ddout ddthree dddifferent ddlanguage ddidentification
ddalgorithms ddto ddsee ddwhether ddeach ddtweet ddwould ddbe ddcategorized ddas ddthe
ddEnglish ddlanguage, dda dddifferent ddlanguage, ddor ddan ddunknown ddlanguage.
The ddfollowing ddtable ddshows ddthe ddproportion ddof ddtweets ddfrom ddeach dddialect
ddthat ddwere ddclassified ddas ddnon-English ddby ddeach ddalgorithm:
AA-alignedWhite-alignedAlgorithm dd113.2%7.6%Algorithm dd28.4%5.9%Algorithm
dd324.4%17.6%
Which ddof ddthe ddfollowing ddstatements ddis ddsupported ddby ddthat dddata? dd- ddAnswer
dd1.All ddthree ddalgorithms ddcontain ddbiases ddthat ddcause ddthem ddto ddmisidentify ddan
ddAA-aligned ddtweet ddas ddnon-English ddmore ddoften ddthan ddthey ddmisidentify ddthe
ddlanguage ddof dda ddWhite-aligned ddtweet.
2.The ddanalytics ddteam ddof dda ddreviews ddwebsite dduses ddthe ddlanguage
ddidentification ddalgorithm ddto dddisplay dda ddpie ddchart ddof ddthe ddnumber ddof ddreviews
ddwritten ddper ddlanguage ddin ddthe ddlast ddyear.
StackOverflow ddis dda ddpopular ddquestion dd& ddanswers ddsite. ddEach ddtime dda dduser
ddasks dda ddnew ddquestion, ddthey ddinsert dda ddrow ddin dda dddatabase ddtable.
Each ddrow ddcontains:
The dduser ddID
The dduser dddisplay ddname
The ddtimestamp ddof ddthe ddquestion
The ddtext ddof ddthe ddquestion
The ddspam ddscore ddof ddthe ddquestion dd(0-5)
Here ddare dda ddfew ddrows ddfrom ddthe ddtable:
user_iddisplay_nametimestampquestionspam_score62038TheAskerator11/27/2012
dd06:15:28How dddo ddI ddgeocode dda ddlat/lng?020394NewCoder12303/12/2015
dd10:55:10Where ddcan ddI ddhost ddmy ddwebsite ddfor ddfree?
136917QuestionErrthing05/04/2014 dd11:34:25Wanna dddownload ddthis ddfree ddfile?3
The ddteam ddwants ddto dddisplay ddquestion ddstatistics ddon ddan ddinternal dddashboard.
, Which ddstatistic ddcan ddnot ddbe ddcalculated ddfrom ddthe ddtable ddof ddquestions? dd-
ddAnswer ddThe dduser ddID ddwith ddthe ddmost ddnumber ddof ddunanswered ddquestions
A ddteam ddof ddscientists ddand ddengineers ddis ddputting ddtogether dda ddresearch ddproject
ddto ddstudy ddwhale ddsounds. ddIn ddorder ddto dddevelop ddthe ddinfrastructure ddfor ddthe
ddproject, ddthey ddneed ddto ddfirst dddetermine ddhow ddmuch dddata ddstorage ddspace
ddtheir ddobservational dddata ddwill ddrequire.
This ddis ddan ddexample ddof dda ddsingle ddobservation:
SoundLocationDate/timeSpecies3 ddminute ddlong ddaudio ddfile63.776871, dd-
171.742193May dd27, dd2019, dd2:23:13 ddPMBeluga
The ddteam ddhopes ddto ddcapture ddthousands ddof ddwhale ddsounds ddfrom ddall ddthe
ddworld's ddoceans.
Which ddpiece ddof dddata ddwill ddincrease ddtheir dddata ddstorage ddneeds ddthe ddmost? dd-
ddAnswer ddRecording ddof ddwhale ddsound
Safiya ddis dda ddsoftware ddengineer ddat dda ddcompany ddthat's dddeveloping ddsoftware
ddfor ddself-driving ddcars. ddShe's ddworking ddon ddsoftware ddthat dduses ddcomputer
ddvision ddand ddmachine ddlearning ddalgorithms ddto dddetect ddpedestrians ddwalking
ddnear ddthe ddcar ddand ddtrigger ddthe ddbrakes ddwhen ddneeded.
After ddtraining ddthe ddalgorithm ddon dda ddlarge dddataset ddof ddtraining dddata dd(videos
ddof ddpedestrians ddwalking ddnear ddcars), ddthey ddtry ddit ddout ddin ddcars ddwith ddbackup
dddrivers. ddThe dddrivers ddreport ddthat ddit dddetected ddmost ddpedestrians, ddbut ddfailed
ddto dddetect ddpeople ddusing ddwheelchairs ddand ddparents ddpushing ddstrollers. ddThe
dddrivers ddhad ddto ddmanually ddrespond ddin ddthose ddcases.
What's ddthe ddbest ddway ddthat ddSafiya ddcan ddimprove ddthe ddmachine ddlearning
ddalgorithm's ddability ddto dddetect ddall ddpedestrians? dd- ddAnswer ddShe ddcan ddadd
ddvideos ddof ddpeople ddusing ddwheelchairs ddand ddstrollers ddinto ddthe ddtraining dddata
ddset dd(perhaps ddcrowd-sourcing ddthem ddif ddthere ddare ddnone ddalready ddavailable).
An ddonline ddcurriculum ddprovider ddoffers ddtheir ddproduct ddto ddtwo ddaudiences:
ddindependent ddlearners dd(self-directed) ddand ddclassroom ddlearners dd(led ddby ddtheir
ddteachers).
They ddwant ddto ddunderstand ddthe dddifferences ddbetween ddthe ddaudiences ddand ddhow
ddthey dduse ddthe ddproduct, ddso ddthey ddsent ddsurveys ddand ddcollected dddata. ddUsers
ddrated ddtheir ddsatisfaction ddwith ddthe ddproduct ddfrom dd1-10, ddwhere dd1 ddis ddleast
ddsatisfied ddand dd10 ddis ddthe ddmost ddsatisfied.
This ddscatter ddplot ddcompares ddthe ddhours ddper ddweek ddspent ddby dda dduser ddto
ddtheir ddrating ddof ddthe ddproduct:
\small{1}1\small{2}2\small{3}3\small{1}1\small{2}2\small{3}3\small{4}4\small{5}5\
small{6}6\small{7}7\small{8}8\small{9}9\small{10}10\text{rating}rating\text{hours/
week}hours/week\purpleE{\text{independent}}independent\greenE{\
text{classroom}}classroom
A ddscatter ddplot ddwith ddhours ddper ddweek ddon ddthe ddx ddaxis ddand ddrating ddon ddthe
ddy ddaxis. ddThe ddx ddaxis ddgoes ddfrom dd0 ddto dd4 ddand ddthe ddy ddaxis ddgoes ddfrom dd0
ddto dd10. ddBoth ddpurple ddand ddgreen dddots ddare ddscattered ddon dd- ddAnswer
