From open data to data-driven services

From Open Data To Data-Driven ServicesMuriel Foulonneau, Slim Turki, Géradine Vidou, Sébastien Martin

ECEG 2014 – 14th European Conference on e-Government12-13 June 2014 - Brasov, Romania

2

Data-driven economy promises

Data-driven economy promises the creation of enormous amounts of economic activity and growth opportunities.

Lie to a large extent in the development of new services.

Results in terms of service creation remain below the expectations of open data promoters.

Most services created are not sustainable and / or do not use the variety of datasets; relying on a limited number of very popular datasets.

June 12, 2014 From Open Data To Data-Driven Services

Open data: Unlocking innovation and performance with liquid information

3

Objectives

Increase the reuse and the value extracted by services from data

Service innovation approaches can help understand mechanisms driving service creation.

1. Review current approaches to encourage the creation of services based on open data

2. Analysis of the creation of services from two open data platforms

3. Identification of the roles that the data can have in the design of services based on a theoretical framework of service innovation.


Approaches to encourage creation of services based on open data

From Open Data To Data-Driven Services 5

Competitions

Opening Datasets

• Datasets available, open licence, sufficient quality

• Interesting for citizens and reusers

Identification of datasets

• Portals aggregate metadata on datasets

Datasets accessibility

• Harmonization of metadata vocabularies (DCAT)

Datasets discoverability

• Lowest possible barrier to reuse

Intellectual Property Rights

June 12, 2014

Publicizing datasets and apps

• Make datasets known to reusers• Events, virtual community channels

Publicizing new datasets

• Section in open data portals to show services and apps using datasets

Publicizing services to show reuse and encourage service usage

Training & support

• World Bank platform, basic knowledge to get data and represent it into maps.

• Open Knowledge Foundation, School of data

Training

• APIs to lower technical barrier and time to develop services

• Reusers do not address format heterogeneity and less knowledge required

SupportHackathons

Off-site competitions

Calls for contribution

Analysis of the creation of services from two open data platforms

7

Observation

Open datasets and services created based on open data listed

Cases selected due to (i) expressed willing of supporting authorities to promote creation of

services based on open data sets

(ii) availability of data concerning open datasets and developed services


8June 12, 2014

315 apps listed• Economy, Environment,

Transparency, Society, Local services, Education, Citizen life

78% reuse only a single dataset• Sketchmap.co.uk uses 19

datasets.

Most reused datasets• Code-Point® Open • National Public Transport Access

Nodes (NaPTAN)..

17869 datasets and 315 apps

102 apps listed• Mobility, Tourism, Environment,

Security, Religion, Business, Open jobs in public administrations.

85% reuse only a single dataset• Only 13 apps use more than one

dataset

Most re-used dataset• “Singapore map (OneMap) data”

used in 33 apps.• Traffic and parking datasets used

in 19 apps.

No more than 100 from 5000 publicly-available datasets are used

From Open Data To Data-Driven Services

9June 12, 2014 From Open Data To Data-Driven Services

Distribution of Apps based on datasets published

Distribution of datasets by topic

Discrepancy between the low number of geographic and transportation datasets and the very large proportion of transportation services.

While transportation datasets appear to have a high reusability potential, education datasets attract less reusers in proportion of their quantity.

Quite uneven potential of datasets for reuse. Domain is clearly not the only element to take into consideration when assessing the

reusability of a dataset.

10

How to increase open data reuse and the number of created services?

Open data portals describe many datasets, available in a variety of formats and with many different access modes

Data producers make their data available for reuse but reuse does not always happen.

Understand the service creation process and how data sources can be integrated in this process.


11

Service design process

Idea Generation

• Ideation phase, birth of the idea• Born spontaneously or from systematic

exploration of various fields of innovation• Can be triggered by a stimulus, call for ideas,

ideation contest.

Maturation

• Exploring the idea related issues• Validate or not options• Investigate which other technologies and

services are out there already.

Concept Evaluatio

n

• Idea has reached a level of maturity, • Assess the potential of the idea by a group of

experts who can decide to invest in its development


Service System

Context

Innovativeness &

Sustainability

SynopsisResources

Target

Synopsis: summary of the concept of the serviceContext (time, space technological components, regulatory context, etc.)Target: customers of the service and the reason why they would buy itResources required (HR, technologies, organization, partners, financial resources…). Service system: way in which resources are combined (key activities, key partners as stakeholders)Innovativeness and sustainability: innovative aspects of the service system, expected impacts.

12

Roles of data in the service design process

Data can play different roles:

1. the service is based on data,

2. the service uses data as a resource, and

3. the service is validated or enriched with data but the data is not directly used or is not directly visible in the service.


13

1- Service based on data (1/4)

When the availability of the data is used as impulse to the service ideation process, it represents the core of the service concept.

• The objective of the ideation process is to determine with one or more datasets which services could be designed based on them.

• The data producer should wonder who currently uses the data; who else could be interested by these data; and if there is any combination of these data with other data which could be of interest to a stakeholder.


14

1- Service based on data (2/4)The service allow visualizing the data

publicspending.net to view budgetary data

nosdeputes.fr, activity of French members of parliaments

handimap.org, paths through cities for disabled citizens

The Narrative science company generates texts from data to make it more user friendly; sport new s articles from the raw results of local competitionsJune 12, 2014 From Open Data To Data-Driven Services

15


Google ngrams benefits from the Google book digitization program:

• Millions of books from various countries at various times digitized;• Optical Character Recognition process to enable full text search

functionalities• By combining the bibliographic data (including the date of publication) and

the individual words used in each book, Google ngrams allows visualizing the evolution of the use of particular words over time.


Evolution of the words "republic" and "democracy" between 1800 and 2000

The service gives a new meaning to the data.

16


Data, main resource or one of the main resources.

Analyse the characteristics of the datasets and their impact on the feasibility of the service:

• update frequency• data quality (reliability, completeness etc.)• data source• maintenance processes, • intellectual property rights attached and conditions of use• cost • accessibility, including its technical accessibility (e.g., API, data dump …) • formats (e.g., RDF/XML, JSON, spread sheet)• interoperability with other datasets, typically to mix it with third party

datasets• documentation including its underlying semantic model to adequately

interpret and use it.


At the maturation phase

17

2- Services with data as resources

Delivery service• Location and traffic data are not the core of the concept.• However they are resources that will help design the service.

Data enrichment• can use Wikipedia for the translation of a dataset of postcodes to

automatically fill the city in an address form.

As for the services based on data, characteristics of the datasets should be analysed


Idea Generation

• Concept defined without any specific relation with the datasets

Maturation

• Investigating service feasibility, datasets have to be taken into consideration as necessary resources

18

3- Service validation

• A dataset of postcodes can be used to validate postcodes provided by users in a form.

• Recommendation systems are often tested against standard datasets: Authors of new algorithms can then test their algorithm against the dataset to verify that it can accurately predict the ratings provided by users.

• Datasets used to validate business models, through gathering economic indicators, from statistical institutes

• Simulation environments require many datasets to recreate the context of execution of a service (traffic related services)

External datasets do not appear in the final service.• Critical role to increase the quality of the service and ensure its viability.


Datasets used in the service design phase but not in the service itself

Concept Evaluation

• use external datasets only for testing a service concept or validating data which are already hold by the service designer.

• Datasets can be used for validating the data already used in the service.

19

Conclusion

Understand the roles that the data can have in a service• Data can help at the maturation and the validation phases of the service

design.• New opportunities for the reuse of data

Different approach to measuring the impact of opening datasets beyond the mere number of services created.

• Benefit of opening data could be adequately measured

Open datasets reuse is not only a matter of promotion but also a matter of asset value, i.e., the characteristics of the datasets are critical to ensure their effective reuse by service designers.


From Open Data To Data-Driven ServicesMuriel Foulonneau, Slim Turki, Géradine Vidou, Sébastien Martin

[email protected] @sl_tu

ECEG 2014 – 14th European Conference on e-Government12-13 June 2014 - Brasov, Romania

Thank you for your attention.

Welcome to Luxembourg

#EDF2015data-forum.eu

November 16-17, 2015

Technology

From open data to data-driven services