Bestand wählen

Using location to inform predictive analytics

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Erkannte Entitäten
lessons will images and politics and acknowledge accompanied by primarily from the about 20 12 contains data science team and we applied in science techniques to on to improve social good in the area of the city engagement health care and education primarily of the examples detail mostly from from politics in the way we go about but the wall of will try to draw the threads into another area that all so basically were all year because of the way we like all circumstantial injured customers to be able to make more decisions on on
really people are actually pretty good at making saga decisions or said that what we really want million scale decisions the too small for a person's and provide people live in the data that they need to make decisions that are appropriately size for Britain to be at the start of the year the example is that you can't run a political campaigns with a human being just looking beachgoers grow violence a cited the 1st persuadable no this 1st lives with the Republicans and a Republican district voted every Republican primary this person is not persuadable you year machine to do that for all of self we work that White decisions can we influence with geospatial the up and what we got to this at the time do we keep tying areas of the answer is not a Julia and that all the way down at the bottom which tourist after everything in between lifeline asking is that the top area that job is all about the analyst decisions that something that a human being said that a sequel prompted census data from the front the budget data sense the datasets this or is that of down at about its sole service tools of work that is either computer is making the decisions for technical users making the decisions based on self tools of so however so that this is actually shifting the areas old specifically to create and sell identifying granularity union unit of analysis of with being with politics that about in health care education lots of places to customer prospective customers to find all the data that you care about people like that of an adult coherent view usually over said but it can be of a building that can be of a neighbourhood that can be any of it can be of road on and then old model that protect the behaviour and that of analysis day strategy based on those things so sure
this is what it looked like off of where you have generally databases which is something like a nowadays verdict that you have the right after a Oracle post grounds and post gases are geospatial a set for supporting a new and that in the right the the ball custom announced exploration tools of south west propensity modeling of you call of a bunch of people and you ask them what they think you figure out what the on what the characteristics are pretty someone support and then you go ahead and that to year during Diouf violent on all every week into a bit better than that that tells you how likely someone is to support your candidate what it doesn't tell you is how likely someone is to be persuaded by the age of certain that to persuade not which we take a set of constituents less than half of delivering message to have those people actually called the other hand someone on are candidate is great for the economy and the other half is actually justice for questioning the use of or persona that person without identifying again the purpose of the 2nd 1 is to make sure they were not just modeling on people were reachable the of so we have to talk to some of real life person at the other end of the following 2 decades ago the person either treatment or a control of the own while they're call up every Monday again after what they think I'm going with 80 see research if the election were held today would you vote for the 1st time that the set up and then could actually is to move based on the message that the recipient versus who it is who who moves with with no method of what really interested in is that there is actually the possibility that some of all this is true in retail to Struan in education through across the board some Fault reaching out actually have a negative persuaded the certainty that we tend to call that sitting down foxes its but after leaving the group on sale would again where they were a geospatial Conference what went interesting is all bunch of those models are are geospatial so 1 example as dry time to avoid location a person is more likely to vote in and out action if if they don't have to drive the authority of the ticket to the polls and so you know someone someone area we did that with a straight radiance just said the number of miles of all and location then spent some time looking at all the other and the drivetime and what is that not all of these look like circles like this is very much not the ball is very much on their even though clusters of a long freeway this was using open straight at data grass and several United just how we think people drive 16 miles an hour on freeway 15 miles an hour on Wednesday for a 20 miles per hour on a number of things around algorithm were used when he now would allow someone to drive out the number of self drive up an offer of rather so it's not an absolute perfect model at a restaurant and a few days building buildings are are now based on the data that we and then validated using and we were like the way you know that takes longer the the set up is so in the end what you end up with is a list of of people with a lot of contact with the ball not the speed of sound were not walk all those talks and the camp not every strategies and individual level of some things happen a gritty returns locations about this is that the sale of the of the density of hours of order we go back and say where she replays offices that are near the their near the most cause of this was the 1st was a from simple really of the The Office that can reach the most people but not to put to the office that could reach the next most people put their Max Mosley for the sale of a further year of the retailer politics all to the good walk through London the systems were really big all year was was how fast and we support as many decisions as possible with the with a reasonably smallest and most of them are there is no going back to the persuasion model thought of all the people who remodelled has been persuadable near by Richmond of turns out there all bunch of those talks wasn't actually worthwhile to rid of walking down really want driveway were at the time and the way that we did that was an iterative where with usual India's area which are a few of the most unsavoury said that the ground everybody within a certain distance of the of the density of with it remains that set of people gravel bunch of people that are with most people and walking out of the blocks somewhere around here in the weather the and the charges in the US was the area these outliers are actually huge apartment buildings the idea of setting up so if 200 people in 1 place it worthwhile to drive the ball in the final on TV advertising of using set about data and location of find the fall to find programs were the most persuadable people what on what your notices there is almost no prime time and where that comes from the name approaches you think people who are interested in politics most likely to use so worried by the end of it turns out everybody advertises news you can get you can get some of tension by advertising in places where the average buying power smaller but the persuasion to sell of set the next part about it is that most of us find it shows were screenshots of you know of an Applications basically everybody who work for the campaign to this is the of this is the Virginia during the Gulf of its way to the Oval with a Democrat still but in the same singles available if you differences version of the storey of walking through a world of the
city's if we can just go browse
so that is what to will you you is
the premiere of the group of this we won the game as simple as possible with a little more salt specify of or a week old let the specified that we call the appointment of 1 of the
region's would of sell basically from the airport to allow them what they wanted to see the correlate got up and ran the ball was in this case were
looking at a density of Judy job by by the Office of by where along with the loss of an emulator was an approach has bought and we make a child's allow Fault to find their own politicians that we have thought of that we also have a new tool that anybody sequels could push lawyers and other they didn't need to be adjusted route the return of worry that attitude and wanted to shut up as if they were in the hands of the high number from 0 2 7 and 8 and some sort of a geographical identify a usually with the type what type of the clothes and and the work of the
of the world where it is that there are Valentia's conduct voters Green Bay Ridge someone means nobody was on the grid and the and the birth of this updated places so focused said pages of the problem here was going on we thought we can trust the somebody whilst to walk or this is a rough neighbourhood we decide the following the of self finally there were obtained the or advanced Alex platforms are really much of a sort of Large Scale and the price of where it was lots of soft where bolted to specific so that you know which we had a sequel training once a week most of oranges musical log on to the database and answer questions using the of a good job of the staff move new to connect or post just over using Yujia baited logging and make their own customer of an so that what is wrong soft were during become 1 of the wrong kind of all systems and area become much more about Coronation of dead data sank into a database document that data build up and provide a piano of provided reinstalls that things like we use your server pretty extensively with players for a a lack of information before walk and tools for pushing the both between the so I've founded by found a group of about India's there really want to find out what it takes to turn from a would push it into a CEO of those longer as a kind of and then tools for pushing data out 2 films that things like explorer of data where she joined showed this was based on the idea of a sequel Guarida shows geospatial they want all my staff to be able to access the so that what all those tools to use in 24 of over here but the or a 150 people on the island extreme in 20 30 array of cells to be able to go back use that faulty just at the time allies in the and we think that means you to be Aizlewood actually means is you need all your data up to date accessible the sequel you a of speed out an essay on the state of the art high or whatever statistical analysis to once you build about all you need somewhere that acted some or over time so I think I'm pretty got out time of so very briefly applying to some other industries primarily health care and education of looking at building the same joint battling infrastructure for those other industries
Nachbarschaft <Mathematik>
Offene Menge
Desintegration <Mathematik>
Service provider
Automatische Indexierung
Gebäude <Mathematik>
Güte der Anpassung
Dichte <Physik>
Dienst <Informatik>
Ausreißer <Statistik>
Rechter Winkel
Einheit <Mathematik>
Physikalische Theorie
Ordnung <Mathematik>
Charakteristisches Polynom
Fluss <Mathematik>
Virtuelle Maschine
Reelle Zahl
Cluster <Rechnernetz>
Statistische Analyse
Abstimmung <Frequenz>
Weg <Topologie>
Kartesische Koordinaten
Fortsetzung <Mathematik>
Strategisches Spiel
Einheit <Mathematik>
Prozess <Informatik>
Große Vereinheitlichung
Zentrische Streckung
Physikalischer Effekt
Abstimmung <Frequenz>
Globale Optimierung
Funktion <Mathematik>
Strategisches Spiel
Extreme programming
Zellularer Automat
Bildgebendes Verfahren
Leistung <Physik>
Physikalisches System
GRASS <Programm>


Formale Metadaten

Titel Using location to inform predictive analytics
Serientitel LocationTech Summit 2014
Anzahl der Teile 14
Autor Dyskant, Erek
Lizenz CC-Namensnennung 3.0 Unported:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen.
DOI 10.5446/15367
Herausgeber LocationTech, Andrew Ross
Erscheinungsjahr 2014
Sprache Englisch
Produktionsjahr 2014
Produktionsort Washington, DC

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract Predictive modeling is used throughout organizations to predict behavior and outcomes; organizations use those predictions to efficiently allocate resources. This talk will cite examples from social organizing and healthcare to show how geographical data can be used to enhance predictive analytics work and drive more efficient and effective programs.

Ähnliche Filme