VII Jornadas eMadrid "Education in exponential times". Antonio Robles, "Using Kibana and ElasticSearch for the Recommendation of Job Offers to Students". 04/07/2017.

1. USING KIBANA AND ELASTICSEARCH FOR THE
RECOMMENDATION OF JOB OFFERS TO STUDENTS
Antonio Robles-Gómez, Salvador Ros, Antonio Martínez-Gámez, Roberto
Hernández, Llanos Tobarra, Rafael Pastor, Agustín C. Caminero, Jesús
Cano
{arobles,sros,roberto,llanos,rpastor,accaminero,jcano}@scc.uned.es
martinezgamez.antonio@gmail.com

2. OUTLINE
 Motivation
 Information Retrieval
 Indicators
 Recommendation System
 Conclusions
2

3.  Current necessity of comparing personal studies with professional
profiles in order to make recommendations about employability needs
 Objective:
 To improve users’ professional career
 Example:
 A user owns a particular degree, but most of his/her jobs of interest require a higher
degree
 It would be desirable to recommend him/her to start a particular post-degree
3
MOTIVATION (I)

4. MOTIVATION (II)
Type Level Who Benefits?
Learning
Analytics
Educational data
mining
Course-level: social
networks, conceptual
development, discourse
analysis, intelligent
curriculum…
Learners, Faculty
Departmental: predictive
modeling, patterns of
success/failure…
Learners, Faculty
Academic
Analytics
Institutional: learner
profiles, performance of
academics, knowledge
flow…
Learners, Faculty,
Administrators
Regional:
comparisons between
systems
Administrators,
Funders,
Marketing
National and International:
comparisons between
National
governments,
Education
4

5.  Several sources of information oriented to the professional market:
 It is essential to extract a common vocabulary in order to make professional
recommendations
 Speak a similar language, independently of the source of information
 Professional social networks:
 InfoJobs, Monster, LinkedIn…
5
MOTIVATION (III)

6. OUTLINE
 Motivation
 Information Retrieval
 Indicators
 Recommendation System
 Conclusions
6

7.  There is two ways of accessing data sources:
 Some of them provides with APIs to access the requested information
 Using crawling/scraping techniques to access the desirable information, due to the lack
some of them are not directly accessible for public usage
 For instance, if you request access to the LinkedIn API, they usually does not allow
accessing its API, in spite of justifying your intentions
7
INFORMATION RETRIEVAL (I)

8. INFORMATION RETRIEVAL (II)
 The organization of information
is not clear and depends on the
particular professional network:
 For this reason, it is essential to
build a common vocabulary, which
includes the most relevant
parameters, in order to make
professional recommendations
 Each offer is composed of a set of
parameters, such as location,
category, minimum degree, years
of experience… 8

9. INFORMATION RETRIEVAL (III)
9

10. INFORMATION RETRIEVAL (IV)
 Elasticsearch for
storing the gathered
data
 Kibana for the
visualization of the
Information
 Discover and graphically
visualize the Information
with panels
10

11. OUTLINE
 Motivation
 Information Retrieval
 Indicators
 Recommendation System
 Conclusions and Future Work
11

12.  Selected indicators:
 Title
 Description
 Duration
 Category and sub-category
 Company name and type
 Location: city/province
 Requirements: experience, degree title, degree title type, and other requirements
 Salary: min-salary, max-salary, currency, and frequency
 Source: Infojobs, Monster…
 Creation date, old and new offers are stored and maintained in the system
12
INDICATORS (I)
Antonio Robles-Gómez, Salvador Ros, Antonio Martínez-Gámez, Roberto Hernández, Agustín C. Caminero, Llanos Tobarra, Rafael Pastor,
Jesús Cano. Defining a Novel Ontology for Educational Counselling based on Professional Indicators (http://educate.gast.it.uc3m.es/wp-
content/uploads/2016/05/uned-paper.pdf). Workshop on Applied and Practical Learning Analytics (WAPLA2016), in conjunction with EC-TEL
2016 Conference. Lyon, France. September, 2016.

13.  Example:
 Java programmer
 This position focuses on programming physical devices
 2 years duration
 Computers and telecommunications categories / Programing sub-category
 Intel company
 Location: Madrid, Madrid
 Requirements: 3 years experience, Master in Computer Science, B2 English
 Salary: 24.000-27.000 euros per year
 Infojobs source
 Creation date: 2016-06-12
13
INDICATORS (II)
Antonio Robles-Gómez, Salvador Ros, Antonio Martínez-Gámez, Roberto Hernández, Agustín C. Caminero, Llanos Tobarra, Rafael Pastor,
Jesús Cano. Defining a Novel Ontology for Educational Counselling based on Professional Indicators (http://educate.gast.it.uc3m.es/wp-
content/uploads/2016/05/uned-paper.pdf). Workshop on Applied and Practical Learning Analytics (WAPLA2016), in conjunction with EC-TEL
2016 Conference. Lyon, France. September, 2016.

14. 14
INDICATORS (III)
Antonio Robles-Gómez, Salvador Ros, Antonio Martínez-Gámez, Roberto Hernández,
Agustín C. Caminero, Llanos Tobarra, Rafael Pastor, Jesús Cano. Defining a Novel
Ontology for Educational Counselling based on Professional Indicators
(http://educate.gast.it.uc3m.es/wp-content/uploads/2016/05/uned-paper.pdf). Workshop on
Applied and Practical Learning Analytics (WAPLA2016), in conjunction with EC-TEL 2016
Conference. Lyon, France. September, 2016.

15.  As a first approximation:
 An educational profile includes the last and previous Degree Titles and Degree Title Types,
and the student’ location
 Possible recommendations:
 A set of offers according to him/her degrees and locations
 Additional studies if he/she wants to reach a set of offers with a higher position, more
salary, or a more prestigious company
 …
15
INDICATORS (IV)

16. OUTLINE
 Motivation
 Information Retrieval
 Indicators
 Recommendation System
 Conclusions and Future Work
16

17. 17
RECOMMENDATION SYSTEM (I)
 Developing a web application in order
to recommend job offers to University
students
 Students will additionally be able to
perform searches according to certain
criteria and offer position features,
such as location, type of company,
duration…
 Visualization panels about job offers
(from Kibana) are integrated in the
web-application

18. 18
RECOMMENDATION SYSTEM (II)

19. 19
RECOMMENDATION SYSTEM (III)

20. OUTLINE
 Motivation
 Information Retrieval
 Indicators
 Recommendation System
 Conclusions
20

21.  Learning does not take place in an isolated context, it is linked with the
professional profiles
 Several professional sources of information, such as InfoJobs and Monster,
have been studied in this work
 A set of indicators are presented for professional profiles
 A recommendation system has been started to guide students (depending of
their type and level of qualification, location…)
21
CONCLUSIONS

22. USING KIBANA AND ELASTICSEARCH FOR THE
RECOMMENDATION OF JOB OFFERS TO STUDENTS
Antonio Robles-Gómez, Salvador Ros, Antonio Martínez-Gámez, Roberto
Hernández, Llanos Tobarra, Rafael Pastor, Agustín C. Caminero, Jesús
Cano
{arobles,sros,roberto,llanos,rpastor,accaminero,jcano}@scc.uned.es
martinezgamez.antonio@gmail.com