Currently working on designing systems and tools to better understand and improve our cities. My main focus goes to Urban Computing, Urban Mobility, Map Inference, and Traffic Analytics. I have previously worked on myriad projects related to social media analytics in the context of #politics, #news, and #ehealth.
I keep publishing every now and then, and act as PC member in several international conferences such as CIKM, CHI, ICWSM, Digital Health, EDBT, and ER.
Education: PhD (Versailles University, France), MSc (Dauphine University, France), Eng (Ecole Superieur d'Informatique, Algeria)
TASMU-QCAI Artificial Intelligence in Transportation Workshop, as featured on Qatar TV
HE Ms Reem Al-Mansoori, Qatar Shura Council Member and MOTCQATAR Assistant Undersecretary for Digital Society Development Sector, and QCRI's Sofiane Abbar are interviewed by Qatar TV at our recent AI in Transportation workshop
[Sept 2019]: Speaking at Qatar University workshop on mobile crowdsourcing for Qatar Smart Cities.
[Jan 2019]: TAWASOL Dashboard received an HBKU Innovation award. -- With Noora Al Emadi.
[Apr 2018]: We started a research collaboration with Land Transport Planning Department @Ministry of Transport and Communication, Qatar.
[Apr 2018]: I'll be speaking at SmartCities conference in Algiers, June 2018.
[Apr 2018]: Our tutorial "The science of Algorithmic Map Inference" is accepted at SIGKDD2018 @London.
[Mar 2018]: Our project "TAWASOL" is accepted to be part of the Innovation Lab @Ministry of Transport and Communication, Qatar. -- With Noora Al Emadi.
[Feb 2018]: Participating in the PC of OpenStreetMap State of the Map conference, Italy, 2018.
[Jan 2018]: Invited to appear in 4 episodes on JeemTV to talk about Computer science, Social Computing, Urban Computing, and more.
[Nov 2017]: Invited by MI department at Algiers' University (La Fac Centrale) to give a talk on Big Urban Data Analytics. Algiers, Algeria
[Nov 2017]: Visiting SISM research division @ Centre de Recherche sur l'Information Scientifique et Technique - CERIST. Algiers, Algeria.
[Apr 2017]: DAWHATI, QCRI’s app to help people better experience and explore Doha during the 2022 FIFA World Cup Qatar, has been short-listed as a finalist for the Challenge 22 competition’s Tourism Experience section. There were just 27 other projects selected as Challenge 22 finalists from an original 937 submissions. With Noora Al Emadi and Jaideep Srivastava
[Oct 2016]: Invited by Sylabs to give a talk on "Smartcities: from data to applications", in the 6th Algiers Innovation Meet-ups, Algiers, Algeria.
QARTA | AI Enabled Routing Engine
QARTA (Carta (it), Carte (fr), Map (en)) is full-fledged traffic aware map engine, delivering several services including: traffic-ware routing and navigation, distance and ETA tables, map tiles, and geo-coding. QARTA is handling million calls per week for large taxi company for operations related to driver dispatching, routing, and fare estimation.
With Rade Stanojevic and Mohamed Mokbel.
Kharita | Building Maps from GPS Traces
Kharita (Map in Arabic) is a robust and online algorithm for map inference from crowd-sourced GPS data. The details of the algorithm can be found in:
Rade Stanojevic, Sofiane Abbar, Saravanan Thirumuruganathan, Sanjay Chawla, Fethi Felali, Ahid Aliemat: Kharita: Robust Map Inference using Graph Spanners. In ACM SIAM SDM'2018 [Arxiv version].
Fork Kharita on Github.
Real-Time Traffic Monitoring and Prediction
We propose a tensor representation for the series of road network snapshots, and develop a regularized factorization method to estimate the missing values, while learning the latent factors of the network. The learned factors, with a graph-based temporal dependency, are then used in an autoregressive procedure to predict the future state of the road network with a large horizon.
With: Abdelkader Baggag, Ankit Sharma, Tahar Zanouda, and Jaideep Srivastava
In IEEE TKDE 2019 [PDF].
ACM SIGSPATIAL 2018 | ETA W-EDGE
Rade Stanojevic, Sofiane Abbar, Mohamed Mokbel. W-edge: weighing the edges of the road network.
Abstract: Understanding link travel times (LTT) has received significant attention in transportation and spatial computing literature. However, the actual LTTs often remain behind closed doors because the data used is considered confidential. In this work we set to enrich the underlying map information with LTT by using a most basic data about urban trajectories, which also becomes increasing available for public use: set of origin/destination location/timestamp pairs. Our system, \wed utilizes such basic trip information to calculate LTT to each individual road segment, effectively assigning a weight to individual edges of the underlying road network.
ACM SIGSPATIAL 2018 | RoadRunner
Songtao He, Favyen Bastani, Sofiane Abbar, Mohammad Alizadeh, Hari Balakrishnan, Sanjay Chawla and Samuel Madden. RoadRunner: Improving the Precision of Road Network Inference from GPS Trajectories
Abstract: Current approaches to construct road network maps from GPS trajectories suffer from low precision, especially in dense urban areas and in regions with complex topologies such as overpasses and underpasses, parallel roads, and stacked roads. This paper presents RoadRunner, a system to improve precision without sacrificing recall (coverage) by proposing a two-stage method.
SocInfo2018 | CityPulse
Sofiane Abbar, Tahar Zanouda, Nora Al Emadi, Rachida Zegour. City of the People, for the People: Sensing Urban Dynamics via Social Media Interactions.
Abstract: Understanding the dynamics of cities is critical towards a better planning of urban strategies. This project is about CityPulse, a system aimed at understanding the complex spatio-temporal dynamics taking place in different neighborhoods within cities. Often, these dynamics involve humans, services and infrastructures, and are observed in different spaces: physical (IoT-based) sensing and human (social-based) sensing.
Tech Transfer: CityPulse Technology is being licensed to CitiesSoft startup.
ACM HT2018 | News Predictions
Sofiane Abbar, Carlos Castillo, and Antonio Sanfilippo. To Post or Not to Post: Using Online Trends to Predict Popularity of Offline Content
Predicting the popularity of online content has attracted much attention in the past few years. In this paper, we propose a new approach for predicting the popularity of news articles before they go online. Our approach complements existing content-based methods, and is based on a number of observations regarding article similarity and topicality. Based on these observations, we use time series forecasting to predict the number of visits an article will receive.