• Near Real-time targeted intervention in high risk population for HIV infection in San Diego County

  • A collaboration effort between Department of CSE and AntiViral Research Centre [AVRC] at UC San Diego

  • Near Real-time targeted intervention in high risk population for HIV infection in San Diego County

  • A collaboration effort between Department of CSE and AntiViral Research Centre [AVRC] at UC San Diego

about us

We employ publicly available data from Twitter to characterize the structure of social networks of local population who may be at risk of acquiring or transmitting HIV infection in the San Diego Area

dashboard-sneak

"Real world assessment of HIV risk is a slow process that relies on static data from national census statistics and other surveys. In some cases this static data is complemented with local data that HIV clinics might have on their patient population, but this only represents the specific subset of HIV at-risk individuals that are already seen at clinics for treatment or testing. However lack of early testing in many HIV at risk individuals, hinders the ability to account for many people who are at risk. Our goal is to reduce this time gap and build a near real-time intervention system to help doctors and researchers to respond to HIV risk effectively. Our work is focused on integrating into a unique pipeline natural language processing, machine learning and network features to filter data publicly available, and confidentially collected from Twitter. The ultimate goal of the developed infrastructure is to help clinicians in characterizing the structure of social networks of local population who may be at risk of acquiring or transmitting HIV infection in the San Diego Area, and to be able to that in real-time. Our integrated platforms allows clinicians to visualize this structure and identify patterns in online social media communication and provide an additional tool to inform targeted interventions to reduce HIV risk."

Infrastructure

Classification

Classify the incoming tweets making using supervised machine learning techniques

Cleansing

Filters the data based on the presence of keywords from the five HIV risk categories.

Adaptation

Improving the vocabulary set based on the inputs from patients at the testing centres.

Remodelling

Re-training the model using updated labels from new set of tweets.

  • "defining network features through interviews and partner tracing would not be as effective"

    Using hiv networks to inform real time prevention interventions

  • "gay and bisexual men accounted for 67% of HIV cases"

    Centers for Disease Control and Prevention

  • "methamphetamine meth ice speed cocaine coke crystal crank"

    Drug Bucket Keywords

Meet Our Team

Project Lead

Associate Research Professor of Computer Science

Dr. Nadir Weibel

Department of Computer Science and Engineering, UC San Diego

Developer

Graduate Student Researcher

Ajay Mohan

Department of Computer Science and Engineering, UC San Diego

Domain Expert

Professor of Medicine

Dr. Susan Little

Department of Medicine, Antiviral Research Center, UC San Diego

Domain Expert

Infectious Disease Expert

Dr. Nella Green

Department of Medicine, Antiviral Research Center, UC San Diego

Former Developer

Software Engineer

Purvi Desai

Apple Inc.

Former Developer

Software Engineer

Narendran Thangarajan

Google

Contact Us




9500 Gilman Drive
La Jolla, CA 92093-0404, USA
pircnet@ucsd.edu