Presentation: How can Plan Ceibal Land into the Age of Big Data?

46
How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France How Can Plan Ceibal Land into the Age of Big Data? Mat´ ıas Mateu [email protected] @Mateu Matias July 23, 2015

Transcript of Presentation: How can Plan Ceibal Land into the Age of Big Data?

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

How Can Plan Ceibal Land into the Age of BigData?

Matıas Mateu

[email protected]

@Mateu Matias

July 23, 2015

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Overview

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Overview

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Overview

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Overview

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Overview

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Overview

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Overview

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Introduction

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Introduction

Uruguay

Uruguay

3.4 M people

GDP per capita 16,400 USD (2013) ranking 44th.

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Introduction

Uruguay

Uruguay’s best well known

Football Association(Soccer)

Asado (Barbacue)

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Introduction

Uruguay

Uruguay’s best well known

Football Association(Soccer)

Asado (Barbacue)

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Introduction

Uruguay

Uruguay’s best well known

Beaches and Resorts:Punta del Este

Yerba Mate

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Introduction

Uruguay

Uruguay’s best well known

Beaches and Resorts:Punta del Este

Yerba Mate

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Introduction

Uruguay

Uruguay’s best well known

Wines: Tannat

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Introduction

Plan Ceibal

Plan Ceibal’s Presentation - Institutional Video

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Motivation

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Motivation

Deployed infrastructure

Laptops and tablets

+700,000 students and professors

Broadband connectivity

+3000 educational facilities with wireless connectivity

Almost 90 % of students with broadband access in schoolfacilities

More than 70 % of students with broadband access athouseholds

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Motivation

Deployed infrastructure cont.

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Motivation

Systems deployed

Platforms

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Motivation

Systems deployed... Generate data

There is a huge opportunity to use it to support strategic decisionsof the Pedagogical/Technological Program in Uruguayan Schools,help measure efficacy of the technologies in hands of students andfinally to develop real time, personalized feedback for students andprofessors to improve and accelerate their learning process

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Motivation

The presented paper

Describes the main data sources, dimensions and variables

Presents some of the challenges and questions that arise totake profit from data

Shows a Case Study

Gives a possible strategy towards implementation of aframework and institutional process for Data Analytics

ID: Data analytics 60119

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Data Sources

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Data Sources

Data Matrix

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Data Sources

Data Generation

Size of Daily Generated Data

Source Size (Mega Bytes)

Zabbix 200CRM 4

Tracker 6PAM activity 10

Internet activity 150

Total 370

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Challenges

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Challenges

Challenges

Great amount of known and unknown variables(hundreds)

Lack of Integration

Lack of a common processing and visualizationframework

Lack of traceability

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Challenges

Challenges

Great amount of known and unknown variables(hundreds)

Lack of Integration

Lack of a common processing and visualizationframework

Lack of traceability

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Challenges

Challenges

Great amount of known and unknown variables(hundreds)

Lack of Integration

Lack of a common processing and visualizationframework

Lack of traceability

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Challenges

Challenges

Great amount of known and unknown variables(hundreds)

Lack of Integration

Lack of a common processing and visualizationframework

Lack of traceability

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Key Questions

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Key Questions

Key Questions

What are the key parameters, significant variables and requireddata sources to include in the integration and exploitation stages?

How can we improve integration of the different data sourcesin a more comprehensive and meaningful way?

How to enable interoperability and consistency betweeninformation and variables retrieved from different data sources(i.e: the unit of analysis in some cases are schools, classroomsor individual based information)?

What are the more reliable analytical techniques to identifystrong correlations amongst key variables?

How can the integration of the different data sources beapplied to better understand ways of improving institutionaland pedagogical strategies?

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Key Questions

Key Questions

What are the key parameters, significant variables and requireddata sources to include in the integration and exploitation stages?

How can we improve integration of the different data sourcesin a more comprehensive and meaningful way?

How to enable interoperability and consistency betweeninformation and variables retrieved from different data sources(i.e: the unit of analysis in some cases are schools, classroomsor individual based information)?

What are the more reliable analytical techniques to identifystrong correlations amongst key variables?

How can the integration of the different data sources beapplied to better understand ways of improving institutionaland pedagogical strategies?

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Case Study

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Case Study

Phase 1

Case Study - First Phase

Motivation

Since 2013 Plan Ceibal has considered the adoption of PAM(Spanish acronym of Adaptive Maths Platform) amongProfessors and students as strategic

Since 2014 Plan Ceibal begun to optimize wifi performance,called High Performace Network (HPN) in every urbanschool facility

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Case Study

Phase 1

Methodology

Hypothesis

1 HPN will facilitate a higher amount of exercises completed bystudents in PAM

2 The social-demographic features (metropolitan vs. interiorurban, and socio-cultural context) affect the use of PAM

Research questions

1 To what extent does network performance correlate with PAMuse?

2 To what extent do the social-demographic features impact therelationships between HPN and PAM use?

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Case Study

Phase 1

Methodology cont.

Key variables

PAM use (number of excercises per student in a given period)

HPN: logical variable

Socio-demographic characteristics (index used at Governmentlevel)

MAC: Assistant Teacher to support use of PAM

Universe and Sample

100 schools with HPN during 2014, 13800 students from 4thto 6th level of primary

Random and stratified by sociodemographic context sample:18 schools with 3,823 students

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Case Study

Phase 1

Preliminary Results

An increase of 35.6% active PAM users has been detected afterHPN was installed

Number of PAM active users before and after HPN deployment

Before HPN After HPN

# PAM active users 806 1093# activities 53179 67523

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Case Study

Phase 2

Case Study - Second Phase

Methodology

A control group was taken: a set of schools without HPNduring 2014

Period of time restricted to second semester of 2014

Factor Analysis based on OLS Multivariate Model to findsignificant impacts in context variables

Dependent variable: Average daily exercises completed inPAM per student

Independent variables:

HPNPresence of MAC (Ceibal’s Assistant) professor in schoolGeographical emplacement of school (urban interior vs.Montevideo)Socio-cultural context of school

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Case Study

Phase 2

Results of multivariate analysis

Factor analysis

All coefficients are statistically significant at p < .05

HPN impact is not significant when it is controlled bysociodemographic contexts

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Case Study

Phase 2

Case study synthesis

In Schools with MAC support, favorable context and urban interior(bivariate analysis: Average Exercices in PAM and HPN) weidentified a significant impact or causal effect. That is to say thatgiven favorable conditions, HPN is something students profit from

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Case Study

Phase 2

Further questions

Technology

Can we find correlations between PAM intensity of use anddevice performance?

Can we find correlations between the use of exercise in PAMand the academic performance of students in Math?

What are the learning outcomes of exercising in PAM?

Can the clustering of teacher’s profile illustrate their influencein PAM’s intensity of use?

Context

To pursue an expanded analysis exploring the impact offactors such as context, geographical location or provision ofTeaching Assistants

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Next Steps

1 IntroductionUruguayPlan Ceibal

2 Motivation

3 Data Sources

4 Challenges

5 Key Questions

6 Case StudyPhase 1Phase 2

7 Next Steps

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Next Steps

Next steps towards Big Data in Education

1 Need for a systematization of the duty of gathering,processing and analyzing data

2 Define targets and Planning (i.e: motivation, engagement,compromise)

3 Create Institutional capabilities to:

design and implement data library and data-warehousegenerate technical skills (measurement and analysis)develop or integrate visualization toolsincorporate decision making processImplement and evaluate periodically

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Authors and Collaborators

Authors

Martina Bailon

Mauro Carballo

Cristobal Cobo

Soledad Magnone

Cecilia Marconi

Matıas Mateu

Hernan Susunday

Collaborators

Helena Rovner

Daniel Castelo

Juan Pablo Gonzalez

Fiorella Haim

Claudia Brovetto

Leonardo Castellucio

How Can Plan Ceibal Land into the Age of Big Data? Data Analytics - IARIA 2015. Nice, France

Thanks! Questions?