Loading...
TICC 01-09-20 (Special) TECHNOLOGY, INFORMATION &COMMUNICATIONS COMMISSION Special Meeting January 9, 2020 7:00 p.m. Conference Room C, City Hall APPROVED MINUTES CALL MEETING TO ORDER Chair Mohanty called the meeting to order at 7:03 pm ROLL CALL Commissioners Present: Prabir Mohanty, Naidu Bollineni, Arnold de Leon, Rajaram Soundararajan,Mukesh Garg Commissioners Absent: None Staff Present Bill Mitchell, Staff Liaison Speakers: Denise Pearl, Strategic Partnerships- Replica Andy Velasquez, Account Executive-Google Cloud Sean Maday, Engineering Manager-Google Cloud Darren Odom, Chief Technology Officer-Boulder Al Nate Newman, Product Manager-Boulder Al Robert Murphy,VP Strategy and Business Dev. -Aclima Joe Hicken, Director Strategy and Business Dev. -Aclima ORAL COMMUNICATIONS This portion of the meeting is reserved for persons wishing to address the commission on any matter not on the agenda. Speakers are limited to three (3) minutes a person. In most cases, state law will prohibit the commission from making any decisions with respect to a matter not listed on the agenda. A. Ilango Ganga,resident B. Zeki Gunam, Cratus Technology, Inc. The Commission reordered the Agenda to hear New Business Item No. 4 first. 1. Receive Google Presentation on Internet of Things (IoT) and Cloud Services. The Commission discussed the Items below: 1. Google Cloud- Smart Communities Overview j • Overview of Data driven decision making, Google's data platform and Google's Approach TICC SPECIAL MEETING MINUTES January 9, 2020 2. BoulderAI-overview of smart cameras/video intelligence • Discuss work with other municipalities 3. Aclima- overview of hyper-local air quality and their work with BAAQMD • San Diego-httns://insi//insights.achma.io/san-dieg_o 4. Replica Overview of platform and urban planning Presentation is attached as part of the record. ADJOURNMENT Chair Mohanty made a motion to adjourn at 9:50 pm., and table the rest of the Agenda items for the February regular meeting. Commissioner Bollineni seconded the motion. The motion passed unanimously. SUBMITTED BY: APPROVED BY: Bill Mitchell, Staff Liaison Prabir Mohanty, Chair Attachment A: Google Presentation PowerPoint I I I I i Attachment A Insights from Data Andy Velasquez ( Account Manager Sean Maday Engineering Manager Google Cloud Google Organize the world's information and make it universally accessible and useful . - Our Mission Government Cloud Use Cases 010 • • ® o �m t) Government Accessibility Employee Productivity Data/Analytics Infrastructure Elasticity Solutions that make Modernizing IT services Leveraging machine Creating an elastic government more open to: and enhancing employee learning and Al for computing framework to productivity data driven decision service new and Limited English making,predictive expanding IT needs Proficient residents analytics and more • Digital Natives meaningful customer • Global Scale • • Google Grade Self-Service interactions. Security • Economies of Google Scale Google Cloud Cp BILLION users You Tube Google in 1 minute - 7. .....-.. .... 1 ,. ��OO r /j l 0 V � ) 1p Tube oo L 0, 1 �a1 010 3M Searches 500 Hours 1000 new ootoV v�j devic 11 es `00111 �00o 1 o ` ` ../�/�< < < <000`Lot O ,,`����x•.`•1K'�. o���a r:Rwne=- r x+:�'a�,� i001 000` o t IL t .1"00101 t !0001 O t t l Q' 401011 l 000 o ! to -90101 1.2M Photos 450K Mile 'fI� ol<<D0 to Oil* Uploaded Driven l0�� OQ61 i 0 011110 Google's Data Research MapReduce TensorFlow Dremel Flume Millwheel Kubernete Borg GFS BigTable Colossus Megastore PubSub Spanner F1 2002 2004 2005 2008 2010 2012 2014 2016 Google Cloud MapReduce Google File System ";;;; �' BigTable w,nr a4uac A amne.ua SMn¢esnlem W Swd,oW Dxu iba Gnopl.Flk Syslem _- _. _ ::... - .. e.uem.,-.nee-mome,.ns+..uwy rr._rr,..n.s W�irae�erw3liw�r�w�rmvy�rE�rrvw mral 18 years of _ EL dvancing the • the ,F 2002 2004 2005 Google Cloud FlumeJava:Easy,Efficient Data-1 MIIIWheel:Fault-Tolerant Stream Processing; The Dataflow Model:A Practical Approach to Balancing Internet Scale Correctness,Latency,and Cost In Massive-Scale, u""""",° 'mm." reradaN.uexam KayaBexuogw,5axacervak. nnab¢r Unbounded,Out-of-Order Data Processing Heaven La:.sam Mcveery Daniel wm,Paw Nadsn¢m..sam vmime IakEau,alengb,lwya0.chernpak haberman, (y:er Aktlau.P¢berl Bratlshsx,daig Chambers,Slava Cbem�ak AWraa `x^w• rNax.ngme,mdlsC,ppn,enm:rehv!@gocgle.cam newel J.FernaMexMOMezuma.P¢urenSam AkVeely,.an el Mills. xynmr.re,n�mLre,..rona... F—Perry,Eric Schmidt Sam Whiffle Wa..puve ._r,y.rYmimrria- AaII'IL\(T m+rvx�auma_sr,sr � w� mbeMb.chemhers.tlrer:ryak rlernand. +T'�—w�a'"`1naO relax,sgmc,ma¢d,lip,cbutle.samualwl@gcogle.com wW pen...ur e AUSTeACl' 1. LYI'NOaUCT10V "'° •' '^" r�n� �.,.+. �� �. t F.,.Nn.mf ru,l 1 atxrll.M1sw�isll�u��w.0 ra..nrm � .erl rxaVM;N�m.mxd. d �I ¢�.lrm.,,.,x.. r. nTanol�cruys •��a�.a ,nka:ne �aL Irp .,n �b wrt. „ _Y.M F x� M W MI ilnv _ u4new.mr ymW.-vs ,M mew aim:w_...x d � �)mr:Nf.,v twRala�4+aekrl .tin � - ��y� _rq 'e'� e...aa.vlxst Mwn.fy e_W:�.Jf,x.trm.+ � - _ r •..'•._-_ N ,.Lwew. „o,.L�rvm,..,w�.,.a...mmre.lx m.m.. �' �..rw .vre�� _ amiymz+� uWtlM.lrM1m ,nvins llife )IV'I,FM 11'-:.o- uLa nm�uu.rrasium+aw+.w n �w"•r rw'i,e .:erewcd+rm�<.au .•�.i„yn-a j.�u.� v"- L�n.M1Mw.h.ibrwN4 x, wrn�.Ip Morn,1e.N ,u xnr,ws.._..iw..,,.rrw smn .x,+mmnesmtM�..+i - u•••vx"'��r�®m' •�`m,W�IdmiaN�mrrumuYY/aN xnm J1 Data 'l* s ver in organizations win or lose based on how they use it Google Cloud Boyd's OODA Loop Observe h: Act Orient Decide Our data analytics design principles CL _ o Focus on serverless Develop integrated Operationalize Iterate and analytics not solutions machine learning measure infrastructure 3 Google Cloud Government Leaders Face Challenges in Delivering Critical Insights Enterprise demand Increasing volumes Data is scattered Security is inadequate to exceeds IT capacity overrun capacity& across multiple protect the organization &funding operating budgets platforms and the mission a ov c' •r �jrrjjj►}i+ir -r 1 - * - � —Ni ' Ci T. Government has valuable data • use Few Organizations • - Organizations Many Organizations Most 3 Organizations ecison akirig 2 Batch analysis ' Ad hoc reporting Machine Learning models Distributed processing actively deployed to Operational frameworks allow increase enterprise Analysts work out of one or organization-wide data to efficiency reporting more data warehouses be used in complex analytic Regularly scheduled used to build reports from processing reports put together by pre-aggregated data hand Standardized Data mining Al/ML Increasing Value to Government Google . • Opportunities Increase operational efficiencies Provide critical cyber - insights Foster civic engagement CloudGoogle Innovators in Transportation V wflvrno Google Maps Waze Waymo Organizations need reliable,real-time Nobody knows what's happening in your Driving today is not as safe or enjoyable as data within user-friendly interfaces to city better than the people who live there. it should be.Waymo's mission is to make it empower their employees and keep Waze exchanges publicly-available safe and easy for people and things to up with their customers'expectations. incidents and slow-down data,enabling our move around.They aim to bring fully Google Maps offers visualization, government partners to respond more self-driving technology to the world that navigation,and analytics to help leading immediately to accidents and analyze can improve mobility by giving people the enterprises chart new territory and find congestion on their roads. freedom to get around and help save transformative new opportunities. thousands of lives now lost to traffic accidents. Google Cloud ' + Roads are busler, heavily, O � congested... ■® ,f =o 0 ...and dangerous Current trends show that by 2030,road The average commuter in metropolitan areas traffic injuries will become the seventh experience 4 hours of road congestion every day.' leading cause of death globally.' 1. US.Department of Transportation,Federal Hlghway Administration Congestion Trends Report I CDC Google Cloud The solution isn't building more roads State and local government construction costs rose 13%in the last five years' d $31.65 . <' It's harnessing your billion data to optimize those � roads 0 0 Google Cloud 1 U.S.Census Bureau,Seasonally Adjusted Data Traffic on 405 Fwy Is Worse 5 Years After$1 Billion Widening Project in Sepulveda Pass: Study PDSTED'd 16-i MA r"1,2019,BY EP C SPA—Ah;,T PACY 2L:0QM AUD 5TE'�E hU_,UILti FED A.] 02'rQPM,MAY Z.019 ®LINKEDIN ('1 PINTEResT Q EMAIL Study 405 Fwy Treffic Worse Afiet Widening Project NEWS Despite a$1 billion widening project to improve traffic along a 10-mile stretch of the 405 Freeway that connects the San Fernando Valley and West Los Angeles,traffic is worse now in the Sepulveda Pass than it was when construction was completed several yeah ago. But that data exists in silos, makingit difficult to use * :— •— r i S 1 LET f Google Cloud Disconnected data is costing us too much 0 C� h� 0 /,; NSC estimates the cost of motor-vehicle More than 40,000 people died in motor deaths,injuries,and property damage vehicle crashes in 2017.2 in 2016 was$416.2 billion. Google Cloud National Safetyof Transportation's 2. US Department of Transportatio!is National Highway Traffic Safety Administration What if your data could be the driver of traffic management? e-0 • Enhance Coordinate Predict Planning Projects Fleet as Data situational response based on road and device urban and optimizing optimizing awareness based real-time insights maintenance transportation transportation assets transportation assets on traffic patterns planning Google Cloud is a cloud-based data analytics platform that brings intelligence, efficiency, and interoperability to CDOT's existing transportation network while enabling world-leading roadway operations for a safer, more reliable, connected, and autonomous future. Google Cloud 22 0 0 0 0 Ingest Transform Store Analyze Interact Get petabytes of data in Prepare,clean,and Create, save, and Derive data insights Explore and from a variety of transform data quickly store datasets at scale without present interactive and formats. and easily. inexpensively. managing servers. impactful data insights. Google Cloud Conversational Interfaces Speech to text queries powered by machine learning What is the average age of the water mains in town? How many homes are on Santa Paula Avenue? =r _e How many bikes per day use the bridge over Stevens Creek Blvd? How many properties added pools last year? Thank you ou r . � I January 2020 What we do Problems we're solving with Cupertino's peers What is Cupertino trying to solve? Boulder I How can city ops and capital keep up with technology? Average Time Spent per Day with TV and Mobile Devices by US Adults,2013-2020 minutes 270 260 250 245 ... ...�.. 238 230 223 229 215 222 219 203 ``� is 188 \,' 170 W0000153 132 2013 2014 2015 2016 2017 2018 2019 2020 IVA ■TV ■Mobile devices Note:ages 18+;time spent with each medium includes all time spent with that medium,regardless ofmultitasking;for example,1 hour of multitasking hour a mobile device while watching ry is counted as 1 hour for Lime _ Nand 1 hour for mobile device (;D Source:eMarketer,April2018 238500 ww+raemerketer.com B I R D 1,Q�e' O �. #FLIR -80% 6 Low detection accuracy $ 1 0k-$20k High price per device Single Purpose i� Siloed object detection capabilities 090%+ ✓` High detection and tracking accuracy Priced to scale Volume pricing for PoC and scaled deployments i Boulder Al DNN Cameras: 0 Many objects • Edge to cloud; no additional Detect,track and count many city objects with one device hardware. • Privacy first sensors. Metadata ® ! } anonymized at the source. • High compute. General purpose deep learning. Continuously improving detection Sensors acquire raw video On-camera analytics extract valuable Boulder Al structures software services metadata from video on metadata from video Pedestrians Near misses 3 Emergency response Cars Intersection timing Bikes Pedestrian safety " Sanitation demand Freight Lane occupancy Public Transport Traffic mix J u Carts HAZMAT Vehicles Scooters Rideshare compliance Parking occupancy Strollers Cyclist/scooter safety • Boulder Al standard • Object detection, counts, • Software (edge or cloud) built hardware platform position and speed data. on metadata or additional (DNN Cameras) • Live and recorded video analytics Making Denver safer is a complicated challenge. ; : DENVER TRAFFIC FATALITIES OVER TIMEIF 70 61 PEOPLE DIED IN 2018 60 • Growth. so 2x population increase since - 1990 (Denver Metro Area) 10 - • Multimodal. 0 Massive rideshare adoption, it changing multimodal fleet. ;;j; 2018 TRAFFIC DEATHS VS MODES • Technology. COMMUTERS Distracted drivers on busier intersections. .% MOTORCYCLESSIKES PEDESTRIANS 3 month milestones: • Determine ped crosswalk occupancy • Count and track peds (anonymously) \ ` '? • Make live data available to signal controller U M •9 = 3 -12,000 v u 8 Q Denver Denver 6,0W _.. 7 Denver O 6 P, COCO -f tip �e to People. 0 01 V4 60% 80% 95% 0 Number of Labeled Images +12000 Combining the best of edge and cloud: t • Search objects, not video. • Explore your data in a w, ­mn. A rich dashboard with real time video, real-time GIS object view. e • Download video clips of important events for context. Camera networks provide metadata Networks of Boulder Al cameras provide private data about objects, vehicles, speed, and position. .,r Layer services on top of metadata !---- ---------C Rideshare cars and Lane occupancy scooters Traffic mix Pedestrian needs Hazards in road Intersection danger Parking occupancy and Emergency response violations IIviDIA. Sanitation service Intersection timing demand HAZMAT vehicles Bicycle safety OHelp Cupertino define a digital infrastructure for operations and growth V 2019ACLINIA INC. CONFIDENTIAL EnvironmentalaCLIma intelligence ,I for people and the planet Aclima Introduction -Cupertino City Council January 2020 Q aWma HEALTH RISK COPD,Asthma,cardiac disease, miscarriage,dementia 92% of the global population FINANCIAL RISK CLIMATE RISK breathes Healthcare costs, ENVIRONMENTAL liabilities,policy risk, property value, 0 RISKQ consumer unhealthy air, liabilities sentiment each day 0 POLITICAL RISK Public sentiment, environmental justice c 2Ci9 ACIMA 111C. CONFIDENTIA aCLIma LimitedWe can't manage coverage Low resolution - - u Pegulatory data today aCUma ■wsni r SIB 3• �.�^� .. r.� ���r��Nafrom" oogle/Aclima -` J ,,� �.7009 arth Q acuma. A unified platform j — - - W" ac�a .ait�i - r I . I 1 2 3 4 5 Mobile& End-to-end network Data Management + Software Tools Data integration for Stationary Sensors management Analytics Intuitive software tools diagnosis+action Best-in-class Unprecedented scale+ Synthesis of Aclima and for experts and citizens Wind,land-use,health data quality block-by-block resolution,via integrated third party to drive action data and more multi-pass driving data to derive insights <-2019 AC.LIMA INC. CONFIDENTIAL Q aCLima. High quality data, lower cost sensors Comparable Performance 09 7WERENCE REFERENCE n —s � High-Cost,Low-Density = O o, oA .................... ..................... 0.! 0.! o.a SENSOR on Low-Cost,High-Density o.t uas tasa uss CONFIDENTIAL Q acuma Backed by scientific f iflo njflL -- - - - encea n A High-Resolution Air Pollution Mapping with Google Street View Cars: rigor at every step Exploiting Big Data Joshua S.dWe.I.a K,ie P.Nesdsr;`uahnd Caul,,agdL,d B_."n�urzus%V.Kirclutcucr, Mdi-N Lundsry''Ju�av D.Alarshilt Chzblophez J.Pod,'Rod CFL Vvm., and S—P.Hamhusgz bR,�men=ra,a,aRees.w>t ma e,�.w�- WP�mq.vd�,o„drotiunmm,n>,m.rm,znzz e�.,axm 'P am�meW Uef RmJ,Yr Yah,.�'ex Yeh ID110 Urild Suue �Sdool d PvpuWlm,M PuWk Uedk Unlmnr d&F.6h CoLm1y Vmmu.v Var 1➢Cmy. Cara aQj 'nd�mr Lr.to zamkw s4 sm P.>nv�m,cWaoml.s51u cow sulm Mellon bryowms d oW as can°m�W v pm�y um mq d W k Yu�w.,e vew5 c ma °IeNmu(s Pxk A,�n�nc4 SJeue,LLaN L'Y•mtq,L'vakl JSN QI.�'tthaLea• 4 � O s,y°wa 1N.�^mb. University STANFORD ---__ __._-- AlSTIACl�AIr Pe:��En-•l1w.d�FY wd1.l1� ,maYmP� m,..Lradf.a3ds► i. . ..o,uu�+a.m mPoa.sm aom• .o+.rl....,M. -IVY � �� E► Maastricht ,���„,` , ��• �P;�� `"" ' r� University " THE UNIVERSITY 3• © ®- 9 P P am,o-,v µ.all..M a♦ 1wy.sw.rw..,-,,. OF Aiuzom. ��' Vi....f,�,..,,n.r,.�.mµ.e.•poem ww Pla6nn w'.P..uaM1.,mPiN^M'..n<I m.50.®av d OaYlmd CA�64'^5'�ISY.e.`■,mler dm d m yR pshW m4 d snud Lmn�'O,\'O�W HNsk oMn u D a..ok mid,ufL,PmYlen P^tum '�ml•Y11b Y �wdde.uu61,),vdnmK Ybdman 4 m Yd wMa ubaidblSYd.Se,x rb P^�•a`mM�•Paa4 4Jk m tea✓^flee.uv�a.J.hm vyaW Igtra., ra kevrr.rus.Ys�m,m m,�.sd,u.....as°•11sa Ir.�aP,eLkk m.®,ypmd mda,aa.. mye�1Jrdr pp sdAY. e 2C19ACUMAINC CONHLIEk-1AL r Q aCLlma Diverse Diversity Experience 40%women expertise i n SO%women in leadership sw ' every domain � zipcar. Diverse team comprised of �' �' 3 _ domain experts from world-class �, I organizations,with a history of building impactful solutions at scale. i rNecuMATE b ICY CORPORATION Bound by a shared mission to build airbnb a global, multi-billion dollar business that drives humanity forward. _ 'I` amazon WORLD BANK M19 AC_I`:IA IN:. CONFIDENTIAL z• Measuring what matters • Carbon Dioxide (CO,) ifi> • Carbon Monoxide (CO) Los Angeles • Ozone (03) • Nitrogen Dioxide (NO,) bi • Nitric Oxide (NO) • Particulate Matter (PMz5) ii • Black Carbon • Methane (CH4) • Total Volatile Organic Compounds (TVOC) <.2019 A:.LI1,1A IN(- CONFIDENIIAL .: _. ..-:... .__<_:....- ::.-�..... o_.m. r.......:...: ........... .__.-� �_ .'r.,i•• yr ��o,�?_*�:..._w��c+�-;1;. Qacuma. Aclima advantage, EXISTING TECH f ro m the user D STEPS months to years) STEPS(-minutes) 1. Determine what to measure,at 1. Subscribe to Aclima Pro appropriate levels ofdetection to • answer user questions perspective 2 Sourcehardware re purchase suitable hardware from multiple vendors 3. Locate appropriate site(s)for continuous monitoring 4. Establish QA/QC plan S. Permit and install hardware, u protecting from the elements Customer problem: I need to 6. Establish cloud based systems and scientific data structures to ingest, understand hyper-local air quality manage,calibrate,harmonize and analyze data rr 7. Hire trained atmospheric scientists to across my city. manage hardware 8. Hire data scientists to interpret results 9. Hire UI/UX designers to visualize large-scale scientific data sets 10. Hire field team to maintain,replace and calibrate sensors 11. Build atmospheric models to translate stationary data into high-spatial resolution outputs Q aCUma EnvironmentalInformed _ Intelligence tools I decisions Aclima SaaS products empower users with best-in-class environmental data Aclima Licensed and analysis tools Softwa re Licensed Aclima software applications are populated with Aclima hyper-local data, in addition to stationary air quality monitoring data from Aclima and 3rd Aclima Data party sources Products 3rd Party Data 2019 AC.UMA INC CONFIDENTIAL Q acuma Aclima Pro 0 Executives and technical staff at a cities and regulatory agencies ask: R What is happening in the vast majority of my jurisdiction that I cannot see? t.7CI9 AC:..INIA I^IC CONFIDENTIAL Q acuma.. Aclima Pro Fo i PF P FF Network operations teams Kj 'V within natural gas utilities ask: Are problematic leaks ' forming in my gas distribution system in between survey cycles? '_Qt9 AC.LIKIA INC CONFIDENTIAL Q acuma. Aclima for Communities Free app for allows the general public to ask: How does air quality impact me and - a my family at our home,and in our community? How can 1 act to make k _ a difference? J N.ub,FlM I a.+WswN.M Par4 ���� �� .g• � /,'4 ,.`�f�s,a�t ^.;<`.:_ CONFIDENTIAL Q acuma Data Products Aclima methodology reveals persistent elevated pollution Annual &Quarterly Baseline levels at hyper-local resolution. Collection results in stable median baseline values for each road segment during the collection period a rc n n Multi-pass Approach E Achieved through repeated driving of target geography Cj L W R Rigorous Statistical Sampling Design Passes distributed across time of day,day of week,and collection period Road Segment Position(km) 2LI19 ACLINIA INC. CONFIDENTIAL Q acuma Google Cloud: Stable, scalable & resilient DATA COLLECTION BACKEND PLATFORM PRODUCTS �I Ingestion&Pre-Processing Aclima Mobile Mapping Data Cloud Storage Cloud Ilr PP g Functions --------------------------------------------- Go HTTP Python Worker BigQuery Aclima Pro Pub/Sub Dataflow Data Lake Go Python Aclima&3rd party Product Store, 0 stationary sensor data Servers&Interfaces Device Data DataProc/ ------------------------------------------------ Hadoop lowQ mapbox 0 Data Analysis TileServer MySQL/Postgres Pipeline Elasticsearch Cloud SQL Front End AN - ------------------------------------------- CS 0 _ ® ' Customer&3rd party data Kb uernetes Container Circlecl Docker NodeJS/React Python Registry Aclima Citizen App ---------------------------------------------I Q acuma Diagnose 0 000 IDENTIFY UNDERSTAND REVEAL INTEGRATE HOTSPOTS& GREENHOUSE EXPOSURE LAND USE SOURCES GASES I I I Act00 , 00 TARGET ENGAGE TRACK OPTIMIZE INTERVENTIONS COMMUNITIES PROGRESS INVESTMENT .2C19aCJllA IN,C CONFIDENI-IPL j t 1 tt r = rm Q acLima Rapidly Expanding Coverage Aclima is contracted with growing number of California public sector users: �t • Bay Area (BAAQMD) • County of San Mateo • LA Metro (SCAQMD) • San Diego County (SDAPCD) • California Air Resources Board (CARB) In the Bay Area,the Aclima fleet is currently deploying region-wide—will map indefinitely Current Aclima deployments i.2019 AC LIMA 011,: High population density areas Q acuma- Cupertino Presentation San Rafael at Aclima Hyperlocal q) `' 4oil '.. a. O, 0 isco t Y.i f� <10 ppb .qp __.. _____ —__..-_ ,; a •\ NO CO C0 > AQMIS Stationary � + San Jos6MountainView Cupertino mapbox ®MaOoox D OpmStrentMap Fe dhn k? Q aWma. Serving Cupertino Aclima is excited to deliver innovative services to enhance community wellbeing and support Cupertino's Climate Action Plan goals through: • Access to hyper-local air quality data and Aclima za,.1 •, Pro analytical tool • Working with city staff to fully understand needs and align product roadmap to meet them • Release of Cupertino-specific public experience • Support from Aclima scientific staff to assess implications of hyper-local data and insights c 2p19 ACUMA INC. COWIDENTIAL Change the wayo lta Yu. - . . see your c *tymove'. 7, F-9 7r 1- B Contldent Rea&&& At the center of THE TRAVEL-DEMAND MODEL many important transportation and land use decisions - SENDA HIRE COLLECT CALIBRATE VALIDATE is a travel-demand PAPER SURVEY CONSULTANTS THE DATA THE MODEL THE MODEL model. Send a paper Hire consultants, Spend 1-2 years Spend 1-2 Certify or validate survey to who will spend collecting traffic years the model 7-10 approximately the next 3-5 and transit counts calibrating years after the The development of a travel-demand 0.5°6 of the years building to calibrate the model primary data model can take upwards often years, region's the model the model was collected population rendering it nearly useless by the time it has been certified. 7' s THE OPPORTUNITY / This creates � ' �0- a significant 1 opportunity to improve the ; r 4 € i ✓.� ;, rµ places we live NINE — N .a by equipping - - public and private1' organizations tl rl I r,a t o .. ; z — W. j with better data and tools. J -- THE OPPORTUNITY R_/___"_ No single source of data is a silver bullet for understanding MOBILE CONSUMER LAND USE/ CREDIT GROUND-TRUTH movement. LOCATION /RESIDENT REAL ESTATE TRANSACTION DATA DATA DATA DATA Replica I Sd—A Labs I Proprietary€Confidential 65 OUR SOLUTION Complete, accurate and representative turn-key solution. MOBILE LOCATION CENSUS GROUND-TRUTH DATA SURVEYS DATA 0 We use mobile We use census We use location data surveys to �f ground-truth to create AI create 100% data including model of travel representative traffic counts behavior synthetic and transit (Personas) population boardings to of each area calibrate our simulation PERSONA SIMULATION& CUSTOMER MATCHING CALIBRATION DEPLOYMENT Al models assigned to Complete internally The model is accessible synthetic households, consistent micro- in Explorer,our preserving privacy simulation of easy-to-use movement querying tool Replies I Sidewalk Labs I Pmprietary€Cmfidmdal 66 OUR SOLUTION / Meet Replica. A shared,fully-calibrated platform for monitoring the movement of people t and goods in a privacy-sensitive way. FRESH We deliver 4 models a year with data for every daytypical I-q u�f�31Irr�..iM►e1� of a week. Id�f�iun,�lgl4 I Ir. �-1rr � .IIIIf1E111U��1�j � t avrl�i r� ` HOW Understand which modes,transit - lines,roads and streets people are using. WHY Explore why people are traveling. WHEN Activities are modeled second-by- second with parcel level precision. WHO Demographic info including income and age while protecting privacy �il�ll VISUAL High-fidelity visualizations of data on the map with choropleths,line maps and more. Rcplic I Sdewnikl.abs I Pmpnemry&Co fidential 67 C 0:0 0UEST10NS-_______ . i 4ITI`ifil p t ` , ;f y -... .. - _.�..•_.�._ Reptim I Sidewalk labs I Pmpdetary BConfidential 68".. QUESTIONS? _= Vy SERE QUESTIONS SWERS