Making Sense of Location Based Services

Locomizer – a solution to Location Based Services’ privacy concerns

The following interview was carried with Dr. Alexei Poliakov and Alexei “Alex” Poliakov, co-founders of Locomizer, on 13th September 2013 at 20 Orange Street, London.

Dr. Alexei Poliakov – a microbiologist with over 10 years of scientific research on spatial pattern formations in mixtures of cells. Alexei is experienced in leading scientific and entrepreneurial teams. His Nature-inspired algorithm is the core of our innovative approach to location data analytics at locomizer.

Alexei “Alex” Poliakov –  holds a strong background in Corporate Biz Dev by building alliances & partnerships in such companies like Fujitsu and Rakuten. He also has experience in hardware & software product planning & management. Alex received Diploma in Computer Info Systems at UC Berkeley & MA in Japanese from FESU, Russia. At locomizer, Alex is responsible for business development/customer development, product strategy/management, and daily operations.

 

With the growth in Location Based Services what makes Locomizer different from the other companies out there?

First of all the following questions must be considered:

Does your company take advantage of location data to improve marketing and if the answer is yes then Locomizer is definitely for you. If the answer is no, then you should really reconsider the use of geo-data because places we visit or come nearby in our real like define our real life interests.

The idea is very simple but significant because “location is our identity”. Location data is generated across multiple verticals like social media, mobile telecoms and credit card payments for example. We all implicitly or explicitly leave digital footprints linked to physical location as we stay and move from place to place. Our strong belief that the era of persistent location is upon us in which the users location data is going to be constantly used to describe, sort and match people based on their attraction and repulsion to real life activities.

The audience discovery engine helps smart companies to take advantage of location data to understand consumer behavior and deliver relevant information.

What is the key USP (Unique Selling Point) for Locomizer?

If a company has location data on its customers, we can build highly targetable user interest profiles by identifying behavioral patterns from their location history. This enables companies to discover and uncover the right audience for the right product and service……

……Simply by answering the question “who to target?”

If companies do not have location data, there is a solution via our geo-behavioral interest graph. An interest graph is an aggregated view of footfall traffic by interest and time interval thus giving contextual knowledge in any given hour, week or month. This helps in being able to make decisions about when and where to target your audience.

Locomizer is backed by more than 10 years research on the behavior of cells. Development of Locomizer’s core algorithm has been inspired by Biology giving an accurate insight to consumer behavior as compared to machine learning because it is based on principles which are 

inherently natural.

That is interesting. What is this “biological algorithm”?

The biological algorithm is based on the way cells behave in space and time. This principle can be applied to explain and predict human behavior – this is the core idea behind the research.

Does this mean that you have identified a correlation or a pattern?

It is not a correlation. The research involves some general rules of science that have been used to understand cell behavior which have been adapted to understand and predict human behavior.

How do you identify an audience for a company?

People make decisions to visit a place which are based on historical patterns as well as needs and desires. Alongside our biological algorithm, research done by other companies and scientists have identified that individuals follow a certain trajectory of movements. This is based on an historical pattern and is not just a random.

We are all creatures of purpose and habit.

If there is a known history of visiting locations Locomizer is able to make accurate predictions rather than guesses. But this is not the core technology. The core of the technology is to take all the personal footprints and place everyone on the same scale (bench mark or ruler).

As an example, the history of three people visiting a location where cinemas are prevalent, like Leicester Square London, can be used as a benchmark. Each person would enjoy visiting this location and do so on a weekly basis but each one would have a different behavioral pattern.

If one of the individuals visits Leicester Square and it is known that he frequents the cinema regularly, his behavior would not be difficult to predict. The opportunities are plenty and the effort made is minimal on the behalf of the individual.

If the second individual has to travel to visit the cinema from a location that does not have one, then this individual would be of interest. He would be visiting the cinema with purpose and would spend time and money to fulfill this need.

In the case of the third individual who frequents Leicester Square because he works in that area there is plenty of opportunity and interest in additional services, such as cinemas and restaurants which can be pushed to him via mobile advertising.

How different is this from Foursquare who get users to checkin to a location?

First of all Foursquare does not offer detailed profiling – what they do is count where people visit and how many times. Foursquare does not take into account the reason behind being in a particular location but only that the person was there – Locomizer takes this into account.

Counting could be equated to rational behavior and we are not rational beings. Foursquare relies on the checkin but for Locomizer it is not a question of checkin but the context of being around a location.

Mobile telecoms collect less accurate location data than Foursquare (within 50 to 150 meter accuracy). Locomizer can work with this and create accurate profiles. 10 meters can make a difference in the profile of a person if using Foursquare but with Locomizer we look at the context which relies on many points. If there is a shift of 10 meters or so due to changes in infrastructure of a location the context remains similar – we can extract the same value from changing circumstances and hence less accurate data.

What actually is provided is footfall traffic by time in a given location which could help retailers, for example, to offer time and location based product promotions.

What about privacy?

The profiles generated don’t tell where individuals spend their time – it is more about how profiles of various sorts are attracted or repelled to/from certain locations by historical activity.

So if an individual visits a place where a cinema is located, that individual would not be that much different than a person who visits the cinema or restaurant in the same location in the same time frame. Both individuals will have separate profiles but “personal information on location, times of visit and where visited” is not gathered or recorded.

As an additional example, a profile may indicate that there is more of an attraction for visiting the cinema than the previous month. It does not indicate how many times the cinema was visited or which cinema.

It gives an indication of a profiles propensity to be around cinemas as compared to the average person who would not visit the cinema in the same time frame of a couple of days, weeks or months.

What about data gathered by mobile companies – how would you deal with that?

We are not data collectors but data processors. We rely on third parties in establishing permission of end users. If the provider wants to store data, consent must be given by the customer.

Locomizer’s is aware of the privacy issue with our aim to protect the consumer. To tackle this we can create accurate profiles without the users ID and telephone numbers. The profiles created are accurate but totally abstract.

Latest Posts

  • RFID, Raspberry Pi & LBS
    The experimentation with Raspberry Pi goes on. RFID is the next stage in effective Indoor Location Based Services and to understand this the following is a small project that brings RFID to the Raspberry Pi. A brief history of the...
    Read more
  • Hacking Location Based Services & IoT with the Raspberry Pi
    Introduction There are very exciting topics that influenced me over the past few years namely Location Based Services (LBS), The Internet of Things (IoT) and the humble Raspberry Pi.  But first a brief background – I am not a technical...
    Read more
  • Arch Linux OpenVPN Server on Raspberry Pi (part 2)
    Once yaourt has been configured, download the “netcfg” package by issuing the following commands; yaourt netcfg There will be a list of programs that have netcfg as a search string. Choose the netcfg by the appropriate number: If you are...
    Read more
LBSZone
Linux Format Magazine

Twitter: global_lbs