May 2016 Paul Lees, group chief executive [email protected] Brian Moran, group company secretary [email protected] Our Greatest Data Hits
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
May 2016
Paul Lees, group chief executive [email protected]
Brian Moran, group company secretary [email protected]
Our Greatest Data Hits
Retained Surplus
£9.9m £18.3m2012 2015
Source: audited Financial Statements 2015
Annual increase in Tenant Arrears
0.07% of rent roll
£43k
Source: Mobysoft March 2016
with half the staff
Annual Void Loss
£404k0.7% of turnover
Source: March 2016 Management Accounts
Non-Emergency Repair Completion Times
7 daysmedian
Source: audited Financial Statements 2015
Net Promoter Score - Tenants
54
Source: quarterly performance report, quart ended 31 March 2016
Net Promoter Score - Staff
47
Source: annual staff survey 2015
Development Programme 2015 - 2018
1,526 units
Source: audited Financial Statements 2015
Housing Management Department (2008)
[THIS PAGE IS INTENTIONALLY LEFT BLANK]
Housing Management Department (2010 onwards)
What’s Going OnMarvin Gaye
10UP 6 1 YEAR IN CHART
“Oh, you know we've got to find a way To bring some understanding here today”
What’s Going OnMarvin Gaye
10UP 6 1 YEAR IN CHART
Rule association
RENT ARREARS RISK
Find the best predictor
OneR algorithm
RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRREEEEEEEEENNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNTTTTTTTTTTTTTTTTTTTTT AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAARRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRREEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAARRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRIIIISSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSKKKKK
FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFiiiiiinnnnnnnddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd tttttthhhhheeeeeeeeeeeeeeeeeeeeeeeee bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbeeeeessssssssssssssssssssssssssssssssssssssssssttttt ppppppppppppppppppppppppppppppppppppppppppppprrrrreeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeedddddiiiiiiiiiiiiiiiiiicccccttttttttttttttttttttttttttttttttttttooooorrrrrrrrrrrrrrrrrrrrrrrrrrrrrFFFFFFFFFFFFFFFFFFiiiiiinnnnnnnnnnnnnnnnnddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd ttttttthhhhheeeeeeeeeeeeeeeeeeeee bbbbbbbbbbbbbbbbbeeeeessssssssssssssssssssssssssssttttt ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppprrrrreeeeeeeeeeeeeeeeeeeeeeeeedddddiiiiiiiiiccccctttttttttttttttttttttttttooooorrrrrrrrr
OnOOnOnOnOnOOneReeeeeReReeReReeeeeeeeReeeeeeeeee a aaaa a a algggglggggggggglgggggggggglggglggggggggggggglgggggggggorrrrrororororrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrorrrrrrrritititititithmhhhmhhhhhhhhhhhhhhhhhhhhhmhhmhhmhhhhhhhmhhhhhhhhhhhhhhhhhh
RENT ARREARS RISK
Find the best predictor
OneR algorithm
VIDEO DEMO
There is a link between missed gas appointments and high (>£250)
rent arrears cases
We can make sense of very large data sets
Let’s Stay TogetherAl Green
9NEW 8 MONTHS IN CHART
“Oh let's, let's stay together Lovin' you whether, whether Times are good or bad, happy or sad”
Let’s Stay TogetherAl Green
9NEW 8 MONTHS IN CHART
K-means clustering
Assign to cluster
K-means clustering
CUSTOMISE SERVICES & INFORMATION
AAssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssiign to cllusssssssssssssssssssssssssssssssssssssssssssssssttttttttttttttttttttttttttttttttttttttttttttttttteer
K-K-K---K-------K--K----K--mmmmmmmmmmmmemmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm ans clusteeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeririrriririrririririrriririirririrrrrrrirrirrirrrriririr nng
CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTOOMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMIIIIIIIIIIIIIIIIIIIIISSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEERRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVICCCCCCCCCCCEEEEES&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&& IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIINNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFOOOOORRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAATTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTIOOOOOOOOOOOOOOOOOOOOOOOOOOOONNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
CUSTOMISE SERVICES & INFORMATION
Assign to cluster
K-means clustering
VIDEO DEMO
We can automatically assign tenants to clusters
We can use clusters to customise services and
information
Robot RockDaft Punk
8NEW 6 MONTHS IN CHART
“Rock, Robot Rock Rock, Robot Rock”
Robot RockDaft Punk
8NEW 6 MONTHS IN CHART
Machine learning classifiers
Check web form
Naive Bayes
Classification
This new case looks like it’s of this type…
TRADE
Classification: Repairs
REPAIR REQUEST
CChheeccckk wweebbb fffoorrmmmCCCCChhhhheeeeecccccckkkkk wwwwweeeeebbbbb fffffooooorrrrrmmmmm
NaNaNaNaNaNNaNNNNNNaNaNaNaaaaaaaaaaaaaiviviivvvve eeeeeeeee ee BaBBBaBBBBaBaaaBaaaayeeyeeyeeyeeyeeyeyyyyeyeyeeeyyeyeyyyyyeyyy sssssssssssssssssssssssssss
CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCClllllllllllllllllllllllllllllaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaasssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssiiiiifififififififififififififififififififififififififififififififififififififififififififififififififificccccccccccccccccccccccccccccccccccccccccccccccaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaatttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiioooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooonnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn
TTThhhiiisss nnnew caase lllooooookkksssTTTThi l kNaNaNaiviviveee BaBaBayeyeyesssNaNaNaiviviveee BaBaBayeyeyesssllllllllllllllllllllllllllllllllliiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiikkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiitttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttt’’’’’’’’’’ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss oooooooooooooooffffffffffffffffffffffffffffffffffffffffffffff tttttttttttttttttttttttttttttttttttttttttthhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhiiiiiiiiiiiiiiiiiiiiiiiiiiiiiissssssssssssssssssssssssssssssssssssss ttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee…………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………
TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAADDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEERRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRREEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAIIIIIIIIIIIIIIIRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR
RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRREEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEESSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
TRADEREPAIR
REQUESTRRREEEPPPAAAIIIRRR
RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRREEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEESSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
Check web form
Naive Bayes
“Our ba th room is leaking through the ceiling onto the stairs and the ceiling is wet through along with the walls where the taps are mounted. And puddle on the stairs”
PLUMBER
Check web form
Naive Bayes
Check web form
Naive Bayes
bathroomleaking
ceilingstairsceiling
wet
walls taps
stairs
PLUMBER
VIDEO DEMO
Machine learning can correctly identify trades for
70%+ of repair requests from tenants
Machine learning can support the creation of
'friction-free' self-service
I Predict A RiotThe Kaiser Chiefs
7
“I predict a riot, I predict a riot. I predict a riot, I predict a riot.”
NON-MOVER 3 YEARS IN CHART
I Predict A RiotThe Kaiser Chiefs
7
Forecasting
NON-MOVER 3 YEARS IN CHART
Forecast time series data
Holt-Winters Triple Exponential Smoothing
FFFFFFFFFFFFFFFFFFFoorrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrreeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeccccccccccccccccccccccccccccccccccccccccccaaaaaaaaaaaaaaaaaaaaasssstttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttt tttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttttiiiiiiiiiiiiiiiimmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmme sssssssssssssssssssssssssssssssssssssssssssssssseeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeerrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrriiiesssssssssssssssssssssssssss dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaatttaaaaaaaaaaaaaaaaaaaaaoooorrrreeeeccccaaaaaaasssstttt ttttiiiimmmmmmeeee sssseeeerrrriiiiiiieeeessss ddddaaaattttttt
Hoolt-Wintters Triplpp eExExpoponenentntiaiall SmSmoooooththiningg
Forecast time series data
Holt-Winters Triple Exponential Smoothing
VIDEO DEMO
We can automatically identify breaks in regular
payments
We can forecast other seasonal data
What Difference Does It Make?The Smiths
6NEW 8 MONTHS IN CHART
“it makes none, but now you have gone,and your prejudice won't keep you warm tonight”
What Difference Does It Make?The Smiths
6NEW 8 MONTHS IN CHART
Randomised Controlled Trials
Test policy
RCT
No SOR provided
Detailed SOR provided
Test policy
RCT
No SOR provided
Detailed SOR provided
Schedule of Rates recording makes little difference to getting the repair done
We can put policy options to the test
The Times They Are A Changing Bob DylanDOWN 2 3 YEARS IN CHART
5
“Come gather 'round people Wherever you roam”
The Times They Are A Changing Bob DylanDOWN 2 3 YEARS IN CHART
5
Crowdsourcing data
"Man is the lowest-cost, 150-pound, nonlinear, all-purpose computing system which can be
mass-produced by unskilled labor."
S. Fred Singer
The Case for Man in Space, 1965
http://quoteinvestigator.com/2016/02/01/computer/#note-12956-10
Crowd sourced data
Adactus500
Crowd sourced data
Adactus500
Apologies to XKCD: see https://xkcd.com/1235/
VIDEO DEMO
Send us a photo of your heating controls
How easy is to to use the controls?
VIDEO
We can get new forms of data from our tenants
Tenants can act as sensors
Communication BreakdownLed Zeppelin
4
“Communication BreakdownIt's always the same”
UP 5 3 YEARS IN CHART
Communication BreakdownLed Zeppelin
4
Data visualisation
UP 5 3 YEARS IN CHART
Data viz
Tableau
Data vizData viz
Tableaaaaaaaaaaaaaaaaaaaaaaaaau
Data viz
Tableau
Data viz
Tableau
VIDEO DEMO
VIDEO DEMO
Risk now has a much higher profile in the boardroom
How information is presented affects the thoughts that
people have about it
Let’s Stick TogetherWilbert Harrison
3
“You ought'a make it stick together Come on, come on, let's stick together”
DOWN 1 20 YEARS IN CHART
Let’s Stick TogetherWilbert Harrison
3
SQL Joins
DOWN 1 20 YEARS IN CHART
Join existing data
SQL
Join existing data
SQL
Join existing dataJ g
SQL
SQL JOINS
Original diagrams: C. L. Moffatt, http://www.codeproject.com/Articles/33052/Visual-Representation-of-SQL-Joins
SELECT <select_list> FROM Table_A A FULL OUTER JOIN Table_B B ON A.Key = B.Key
Outer (Full) JOIN
SQL JOINS
Original diagrams: C. L. Moffatt, http://www.codeproject.com/Articles/33052/Visual-Representation-of-SQL-Joins
SELECT <select_list> FROM Table_A A INNER JOIN Table_B B ON A.Key = B.Key
Inner JOINSELECT <select_list> FROM Table_A A FULL OUTER JOIN Table_B B ON A.Key = B.Key WHERE A.Key IS NULL OR B.Key IS NULL
Outer Excluding JOIN
SQL JOINS
Original diagrams: C. L. Moffatt, http://www.codeproject.com/Articles/33052/Visual-Representation-of-SQL-Joins
SELECT <select_list> FROM Table_A A LEFT JOIN Table_B B ON A.Key = B.Key
Left JOINSELECT <select_list> FROM Table_A A LEFT JOIN Table_B B ON A.Key = B.Key WHERE B.Key IS NULL
Left Excluding JOIN
SQL JOINS
Original diagrams: C. L. Moffatt, http://www.codeproject.com/Articles/33052/Visual-Representation-of-SQL-Joins
SELECT <select_list> FROM Table_A A RIGHT JOIN Table_B B ON A.Key = B.Key
Right JOINSELECT <select_list> FROM Table_A A RIGHT JOIN Table_B B ON A.Key = B.Key WHERE A.Key IS NULL
Right Excluding JOIN
VIDEO DEMO
We can create balanced measures of staff
performance
We can connect together many sources of data
If 6 Was 9The Jimi Hendrix Experience
2
“Now if a 6 turned out to be 9 Oh I don't mind, I don't mind”
UP 1 20 YEARS IN CHART
If 6 Was 9The Jimi Hendrix Experience
2
Feature creation and engineering
UP 1 20 YEARS IN CHART
Create new feature
SQL
Create new feature
SQL
TENANCY BALANCE
REPAIRS SPEND
RECHARGEABLE REPAIRS
ASB CASES / LEGAL ACTION
GAS NO ACCESS
COMPLAINTS
CRM CONTACTS
PARETO SCORE
Distribution of Pareto Scores
Distribution of Pareto Scores
0.7% of tenancies are challenging
Less than 1% of tenancies are challenging to manage
We can figure out ways to measure proxies for data we
don't record
SurvivorDestiny's Child
1NON-MOVER 2 YEARS IN CHART
“Even in my years to come I'm still gon be here'”
SurvivorDestiny's Child
1NON-MOVER 2 YEARS IN CHART
Survival analysis
Survival Analysis
Kaplan-Meier Estimator
S(t)=1−F(t)= Prob(T≥t)
SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuurrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvviiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiivvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvaaaaaaaaaaaaaaallllllllllllllllllll AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAnnnnnnnnnnnnnnnnnnnnnnallysiissssssssssssssssssssssssssssssss
KaKKKaaaKaaaKaKaKKaaaaKaaKaKaKaaaaaaaKaKaaaKaaaaaaaaKaKaaaaaaKKKaaaKaKaKaaaaaaaaaaaaaplpplplpllppplplpplplplplpllppppplllllpplppllpllpppplllpllplllplplppppllpppppllpllpplplllllllpppppppllpppplp ananananaaananananaanannanananannanaaaaaanaaaaananannnnannnananaaaanaananannananaaaaaanaaanananannnanaaananaananaaaannaanaaaaaaannnnnaannaaaaannnnnnnannnaanaaaaaaaaanaaannnnnanaann-M---M-M-M-MMMMMMM-M-M-M-MMMMMMMMMM-MMMMM-M-M-MMMMMMMMMMMMMMMMMMMMMMMMM-MMMMM------MMMMM--MMMMMMMMMMMMMMMMMeieieieieieieiieieieieieieieiieieieieieiieieeeeiiiieieiieieiieeiiiiieieiieieeeeieieieieieieieieeeeeiiiiiiiiiiiiee ererereeeeereerereeerereeeeerereeeeeerereeeeeerrreeeerereeeeerererrerrereeeeeeeerrrrrrereeerrrererreeeeerer EEEstimattttttttttttttttttttttoroo
SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((ttttttttttttttttttttttttttttttttttttttttttttt)))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))===========================================111111111111111111111111111111111111111111111111111111111111111111111−−−−−−−−−−−−−−−−−−−−FFF(((ttt)))))))))))))))))))))))))))))))))))))))))))))))))))))))============================================================================================ PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooobbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb((((((((((((((((((TTT≥≥≥TTTTT ttttttttt))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))
Survival Analysis
Kaplan-Meier Estimator
S(t)=1−F(t)= Prob(T≥t)
VIDEO DEMO
Histogram of Tenancy Length (Current Tenants)
Survival Analysis can be applied to many other longitudinal data sets
Most tenancies churn quickly
Housing Association Costs
A Simple Model of Tenancy Lengths
Tenancy Length (years)
Tenancy Length (years)
A Simple Model of Tenancy Lengths
Tenancy Length (years)
Unlikely to need component
replacements
Likely to needcomponent replacements
Implications for Planned Maintenance
Implications for Planned Maintenance
The 60:40 Rule is inefficient
Implications for Planned Maintenance
Major Works & Cyclical Maintenance Spend
£888pu £679pu2012 2015
Excluding discretionary spend
Rule association
K-means clustering
Machine learning classifiers
Forecasting
Randomised Controlled Trials
Crowdsourcing data
Data visualisation
SQL Joins
Feature creation and engineering
Survival analysis
May 2016
Paul Lees, group chief executive [email protected]
Brian Moran, group company secretary [email protected]
Our Greatest Data Hits
Related Documents
