In recent years, geolocalisation technologies have created "Location Based Services": products and services adapted to people's past and present geolocations. Made possible thanks to Roofstreet technology, the next evolution is in progress with the arrival of "Trip Based Services": products and services that adapt not only to positions but also to past and future trips.
Who needs to understand travel today?
Understanding travel is a current issue and essential to many sectors.
While Retail now uses a lot of offlines data from, for example, loyalty cards or some point-of-sale information (postal code), onlines data are more difficult to find and are mainly used by third-party players in the market. mobile advertising for mobile targeting and visitation purposes.
Retailers must also make decisions on display strategies, implementation of point of sale, or understand their position relative to the competition in terms of attendance on a particular consumption basin.
Retailers offer to differentiate more and more services to their customers in order to simplify their lives. The digital vector, application and mobile site is often the channel allowing them to make a difference.
Thanks to mobile applications and sites, it is possible for them to better know their customers (products consulted, intentions to purchase), but we note that the use of geolocation in the context of customer knowledge is still under exploited.
Media agencies and some retailers are building specific departments in the analysis of travel behavior given the importance of the challenge for the future.
This diagnostic phase must identify the population movement flows in order to plan and optimize the sizing of the infrastructures and to plan the routes of the various available lines as well as possible.
It also makes it possible to supplement the lack of transport supply in certain areas with alternative modes of transport that are either soft or shared (carpooling).
In a more operational logic, the carriers seek to provide quality services to users and in particular by setting up proactive services, anticipating the movements of these to know them better and inform them specifically about their journeys and only at the moment. timely.
Moreover, challenges remain on the ability to trace the entirety of a user journey, if the ticketing data allow most of the time to detect the entry into a mode of transport modulo the fraud, they do not allow on the other hand in most cases to know the exit places of these users.
The available flow data are usually presented from one municipality to another with a count of the number of people making a commute to work over a year.
Dividing the territory into 16,000 "IRIS" zones, each zone gathers on average between 1800 and 5000 people according to the densities of the agglomerations and have an average surface area of 41KM².
Most geomarketing tools and feed data providers use the IRIS field as the most granular area for representing origin / destination matrices.
This use presents several difficulties, on this type of data it is for example not possible to obtain the flow per borrowed axis, nor to go below the commune or the IRIS zone, nor to consider a period of time finer than that studied at the orgine.
The area of the IRIS areas is sometimes too large to respond to fine analyzes, especially in dense urban areas.
Finally, customers often have their own areas of analysis that do not always coincide with the proposed areas.
Many companies use field surveys conducted by specialized institutes: these surveys can be carried out by physical counting or by telephone.
Field counts are usually performed on 1 or more specific areas requested by the customer. A count covering the whole of an agglomeration would be too expensive.
Some of them make it possible, in particular, to reconstruct the pathways of panelists on a declarative logic.
If the statistical extrapolation models are very robust, the samples taken are often very small.
Tracers can be either GPS or specific boxes that collect all the information on the conduct of the person. The data provided is of good quality and can also be used to understand driving behavior (useful for example to insurers).
Devices of this type are generally hardware, they require the acquisition of hardware, their installation and uninstallation as well as the associated software and hardware updates.
Some mainstream applications (such as Waze) can calculate actual road routes, but these data are the property of those who collect them and are not available on the market.
In addition, these data only cover motorists and therefore do not allow for non-motorized or public transport trips.
Public transport data:
Public transport open data if they can be very useful to know the bus schedules, the stations, routes of lines, station of self-service bicycle, concern very little the displacements of people.
The ticketing data attempting to model the routes takes into account only the entries in the vehicle but not the exits.The user does not validate his ticket at the end of the trip so the reconstitution of the journey is only partial.
Telecoms typically provide origin-destination matrices based on the location of people in areas of telephone or aggregated antennas at IRIS.
If the representativeness is interesting given the penetration rates and market share of each actor, technical limitations exist and generally do not allow to go down to a level of precision less than several hundred meters.
Data from GPS tracks of smartphones
This data comes from embedded location systems in smartphones.
SDKs (Software Development Kit) are integrated in consumer mobile applications and collect (if the user has given them the right) the positions returned by smartphones.
These data are used today primarily for advertising and customer targeting purposes by geofencing methods or segment targeting of people who have been seen at a location that is of interest to an advertiser.
These data are of great value because of their precision which varies between a few tens and several hundreds of meters, and by their frequency of collection and their real time aspect.
There are, however, common geolocation errors (use of underground transport) and accuracy is not always proven. In addition, the frequency and the mode of collection of points becomes more and more restrictive and the collected data are more and more fragmented.
Actors processing this type of data must ensure that they have implemented algorithmic methods that detect errors, delete or weight the data according to their accuracy.
While this data may be relevant for mobile advertising stakeholders in order to target and measure visits to a point of sale, it is difficult to use to model a person's journey in its entirety and does not allow us to understand where did a person pass between two point surveys.
Trip-Based Services: the paradigm of "going through" / " will go through":
Smartphone data is highly accurate notwithstanding the ability to filter, process and correct geolocation errors.
Roofstreet has developed algorithms that allow on a weak base point to reconstruct all of a person's journeys.
The algorithm will thus link the points between them in an intelligent way by observing the habits.
The more repetitive the motion, the more accurate the path returned by these algorithms.
Moreover, when a trip is created it is possible to understand how a person moves from point A to point B in order to detect recurrent displacements and to use them as a basis for predicting the next trip (s).
The ability to generate precise paths on a small number of points collected makes it possible to overcome major constraints being:
Save the phone's battery consumption
Respect the restrictions imposed by manufacturers on the frequency of collection.
Cases of use made possible by the knowledge of the journeys
Thanks to the creation and aggregation of individual paths, it is now possible:
- For local authorities, to know the flow of movement by Road or non-road between neighborhoods and / or communes / generating poles and with a very fine temporal granularity (time, typical day, typical week).
- For the retail sector, catchment area 2.0 consists of identifying the provenances, places of residence or work and the journeys made by any person before going to a given area. Only paths can reconstruct this type of history.
Retailers can now rely on a database of real-time refreshed travel data to optimize marketing investments, particularly with regard to urban signage thanks to audience measurements, to define the ideal location for the positioning of customers. long-life panels as well as the best distribution areas for advertising printed matter in combination with existing data sets.
In addition, the e-commerce players are currently considering a way to take into account the journeys of a person to offer him adapted withdrawal points to facilitate his experience with the brand.
- For the transport, the knowledge of the paths associated with the line plan makes it possible to know automatically, without having to ask the user, which lines it borrows in order to be able to warn it in advance, on these lines and only at the right time, disturbances.
From an operational point of view, these paths and travel intentions can be provided to the operators of the control towers to provide an additional decision element for the programming of train departures for unloading on the platforms.
Moving from "location based services" to "trip based service" is therefore an issue with great potential for any brand or public service wishing to better understand travel and communicate better with its customers.
Source: Roofstreet.io Blog - 11 Jan, 2018