3.7. Resonance API Client¶

3.7.1. Theoretical Basis¶

Information about the activities listed in Monitoring Events and the snapshot of the device containing geographical data of itself are uploaded on our server for storage and periodical analysis in order to identify:

• Recurrent locations - The system can identify the coordinates of the user’s home and job defining as home “the place where the user sleeps” and as work “the major location where every-days activities are done”.
• User’s routine - The algorithms developed estimate the probability that the user is performing a specific activity at a specific time.
• Points of interest - Restaurants, cinemas, theaters, museums and so on, according to user’s locations.

To process all these data we use temporal series analysis strategies as: study and evaluation of chaotic attractor regarding users movements, determination of maximum Lyapunov exponent [ Wolf1985 ], [ Mukherjee1997 ], [ Franchi2014 ] and clustering algorithms as K-means and G-means.

3.7.2. ResonanceApiClient class¶

ResonanceApiClient is a singleton class that should be started together with Android Application.

It follows common logic to be adopted for a Resonance ready application:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

public class ResonanceApp extends Application {
  @Override
  public void onCreate() {
      super.onCreate();
      // building ResonanceApiClient instance
      ResonanceApiClient client = ResonanceApiClient.with(this);
      // starting ResonanceApiClient instance
      client.start();
  }
}

In reported code, logic for initializing ResonanceApiClient instance is encapsulated into with() method, that is responsible for defining several parameters in several areas (e.g. parameters for activity tracking, endpoints base url and so on). Nevertheless it’s possible on client side to exploit ResonanceApiClient.Builder class for manually define ResonanceApiClient configuration, but this practice is discouraged in most cases, since basic settings are commonly adequate for most cases.

3.7.3. Data Collector¶

When starting ResonanceApiClient instance, ResonanceSDK automatically starts collecting data concerning device usage. Data are periodically delivered to server with adequate strategy for minimizing battery consumption, and allow to perform analysis on user behavior, determining relevant information about his preferences. This is essential for making predictions as well as for identifying changes in his regular behavioral pattern.

Going in depth with description of data collector is not essential in this documentation, since developer doesn’t need to interact with it in any way.

3.7.4. Advisor¶

Outcome of data analysis performed by Atooma backend is made available through a set of useful functions implemented by a dedicated class that is ResonanceAdvisor. Below is shown code for getting such instance:

1
2

Context context = getApplicationContext();
ResonanceAdvisor advisor = ResonanceApiClient.with(context).getAdvisor();

1
2
3
4
5
6
7
8

// Returns a list of locations visited by the current user / device
void getRecurrentLocations(AdvisedElementsResponseHandler<AdvisedLocation> listener);

// Returns a list of locations matching home type, visited by the current user / device
void getHomeLocation(AdvisedElementsResponseHandler<AdvisedLocation> listener);

// Returns a list of locations matching work type, visited by the current user / device
void getWorkLocation(AdvisedElementsResponseHandler<AdvisedLocation> listener);

1
2
3
4
5
6

advisor.getRecurrentLocation(new AdvisedElementsResponseHandler<AdvisedLocation>() {
  @Override
  public void onAdvisedElementsRetrievedListener(List<AdvisedLocation> locations) {
    // work with locations here
  }
});

1
2
3
4
5
6

advisor.getRestaurants(new AdvisedElementsResponseHandler<AdvisedRestaurant>() {
  @Override
  public void onAdvisedElementsRetrievedListener(List<AdvisedRestaurant> elements) {
    // execute code
  }
});

1
2
3
4
5
6
7

advisor.getRoutine(ResonanceAdvisor.ROUTINE_TYPE_MONDAY,
      new AdvisedElementsResponseHandler<AdvisedRoutineItem>() {
    @Override
    public void onAdvisedElementsRetrievedListener(List<AdvisedRoutineItem> elements) {
      // execute code
    }
});

3.7.5. Next Steps¶

A new system that harvest information from other services (Facebook, Twitter, Openstreetmap, Tripadvisor) and wearable devices is currently at study. It will improve the recognition of the activity currently performed and will enable us to develop suggestion and tips for users regarding their activities: have dinner, see a movie, listen to a concert and so on.

3.7.6. Bibliography¶