Thread by @NAChristakis

Today, we are happy to release a COVID19 app that is unlike any others available.

HUNALA allows you to track & FORECAST your personal risk for respiratory disease daily. It’s like traffic apps (eg Waze) for coronavirus & respiratory disease. Get it at hunala.yale.edu 1/

Users contribute information anonymously and get aggregated risk predictions, just like Waze forecasting traffic jams miles ahead on a highway. The data users contribute is statistically combined with public information about respiratory disease and COVID19. 2/

The more people who use HUNALA overall (get it here: hunala.yale.edu), and the more who use it in any particular area, the better it gets. HUNALA is released by @yale and by HumanNatureLab.net. #HNL 3/

We pronounce HUNALA who-NA-la. Just saying.

And its origin story involves a chance dinner @yale in February with @alexichristakis. A large team (tagged below!) has been working ever since then. Don’t ask why it took so long to get to launch. 4/

HUNALA is not a citizen science app, collecting info for use by other actors (though it's indeed pro-social to use it). It is not a contact tracing app, where a user is told about PAST exposures.

HUNALA is a forecasting app that gives users info about their own FUTURE risk. 5/

HUNALA does not give medical advice. 6/

HUNALA respects privacy and anonymity. Recently, @mike_giglio wrote about how our approach makes sense in this regard in @theatlantic theatlantic.com/politics/archi… 7/

HUNALA asks a few questions about you on enrollment. Thereafter, you answer a couple of questions a day if nothing is happening, and you answer less than a minute of questions if you have symptoms or have seen a doctor since your last check-in. 8/

HUNALA also collects some information about your real social network using your contacts. This is anonymous. Your contact list is not copied or shared. People you identify are pinged just once and told about HUNALA, if they want. 9/

Every time you use HUNALA, you get a personalized machine-learning-developed forecast of your risk for contracting a respiratory disease based on where you live and based on where you are in the social network (and also based on other measures). 10/

You can use location tracking (only used ONCE a day, when you input your information – it is not on in the background!) in order to get an assessment of COVID19 conditions for where you happen to be at the time, if you are traveling or away from home. 11/

Soon, we will enable users to assess risk in any county in the USA, for instance before traveling or in order to monitor conditions in locations where a friend lives – like a weather app. We will soon be posting maps daily on hunala.yale.edu. 12/

Just like traffic apps, all the information is anonymous.  We do NOT inform anyone about anything happening to you personally. Nor vice versa. 13/

HUNALA will be useful as people emerge from lockdowns and during the likely second wave of the COVID9 pandemic excepted to strike the USA in the fall of 2020. It is also useful for colleges, hospitals, factories, workplaces. 14/

The app is based on network science principles, including partly on work we did with the H1N1 pandemic in 2009. journals.plos.org/plosone/articl… 15/

Some of the network science principles (e.g., on using networks to forecast epidemics) which are embedded in HUNALA are also described in this @TEDtalks ted.com/talks/nicholas… 16/

Other science we have published – on privacy protection, machine learning, and network science – relevant to HUNALA includes:

iid.yale.edu/publications/2… &

iid.yale.edu/publications/2… &

journals.plos.org/plosone/articl…

| @aminkarbasi 17/

The forecasting part of HUNALA involves a cool new machine learning component spearheaded by @aminkarbasi & @jacob_derechin. The more people who use the app, the better the forecasts will get. In this way, HUNALA resembles traffic apps. 18/

Key members of the @yale team working on this include @d_i_s_p_e_r_s_e, Wyatt Israel, @AlexiChristakis, @aminkarbasi, @jacob_derechin, @ShivkumarVs, Laura Forastieri, @TKeeganT, @EmilyDeegan | @YINSedge 19/

For an FAQ about the HUNALA app that allows personalized forecasts for risk of COVID19 and other respiratory diseases, given where you live and what is happening in the network around you, using cool network science and machine learning, see: hunala.yale.edu. 20/

For other network sciences software (used for research purposes) previously developed and released by the Human Nature Lab #HNL @yale @YINSedge , see:

trellis.yale.edu

breadboard.yale.edu

| @d_i_s_p_e_r_s_e 21/

Please send suggestions for features (that are realistic!) and bug reports about HUNALA by responding to this tweet. We will announce new features in the coming months by adding to this thread. | @EmilyDeegan 21/

HUNALA, our app for forecasting personal risk of respiratory disease, is not yet available for use in countries outside the USA. We are working on it. 22/

Today, we're pleased to release a new service, as part of our #HUNALA app (an app that individuals can use to track their own personal COVID19 risk): it's an interactive daily map of county-level risk. Map (scroll down) and app (if you want it) are at: hunala.yale.edu 23/