BEGIN:VCALENDAR VERSION:2.0 PRODID:-//pretalx//talks.osgeo.org//foss4g-2022-academic-track//SMRLLD BEGIN:VTIMEZONE TZID:CET BEGIN:STANDARD DTSTART:20001029T040000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:CET TZOFFSETFROM:+0200 TZOFFSETTO:+0100 END:STANDARD BEGIN:DAYLIGHT DTSTART:20000326T030000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:CEST TZOFFSETFROM:+0100 TZOFFSETTO:+0200 END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT UID:pretalx-foss4g-2022-academic-track-VMNCM3@talks.osgeo.org DTSTART;TZID=CET:20220825T141500 DTEND;TZID=CET:20220825T144500 DESCRIPTION:Motivation:\n\nBecause of technological advancements\, public p articipation in scientific projects\, known as citizen science\, has grown significantly in recent years (Schade and Tsinaraki 2016\; Land-Zandstra et al. 2016). Contributors to citizen science projects are very diverse\, coming from a variety of expertise\, age groups\, cultures\, and so on\, a nd thus the data contributed by them should be validated before being used in any scientific analysis. Experts typically validate data in citizen sc ience\, but this is a time-consuming process. One disadvantage of this is that volunteers will not receive feedback on their contributions and may b ecome demotivated to continue contributing in the future. Therefore\, a me thod for (semi)-automating validation of citizen science data is critical. One way that researchers are now focusing on is the use of machine learni ng (ML) algorithms to validate citizen science data.\n\nMethodology:\n\nWe developed a citizen science project with the goal of collecting and autom atically validating biodiversity observations while also providing partici pants with real-time feedback. We implemented the application with the Dja ngo framework and a PostgreSQL/PostGIS database for data preservation. In general\, the focus of biodiversity citizen science applications is on aut omatically identifying or validating species images\, with less emphasis o n automatically validating the location of observations. Our application's focus\, aside from image and date validation (Lotfian et al. July 15-20\, 2019)\, is on automatically validating the location of biodiversity obser vations based on the environmental variables surrounding the observation p oint. In this project\, we generated species distribution models using var ious machine learning algorithms (Random Forest\, Balanced Random Forest\, Deep Neural Network\, and Naive Bayesian) and used the models to validate the location of a newly added observation. After comparing the performanc e of the various algorithms\, we chose the one with the best performance t o use in our real-time location validation application.\n\nWe developed an API that validates new observations using the trained models of the chose n algorithm. The Flask framework was used to create the API. The API uses the location and species name as parameters to predict the likelihood of o bserving a species (for the time being\, a bird species) in a given neighb orhood. Moreover\, the model prediction\, as well as information on specie s habitat characteristics are then communicated to participants in the for m of real-time feedback. The API has three endpoints: a POST request that takes the species name and location of observation and returns the model p rediction for the probability of observing the species in a 1km neighborho od around the location of observation\; a GET request that takes the locat ion of observations and returns the top five species likely to be observed in a 1km neighborhood around the location of observation\; and a GET requ est that returns the species common names in English.\n\n\nUser experiment :\n\nA user experiment was carried out to investigate the impact of automa tic feedback on simplifying the validation task and improving data quality \, as well as the impact of real-time feedback on sustaining participation . Furthermore\, a questionnaire was distributed to volunteers\, who were a sked about their feedback on the application interface as well as the impa ct of real-time feedback on their motivation to continue contributing to t he application.\n\nResults:\n\nThe results were divided into two parts: fi rst\, the performance of the machine learning algorithms and their compari son\, and second\, the results of testing the application through the user experiment.\n\nWe used the AUC metric to compare the performance of the m achine learning algorithms\, and the results showed that while DNN had a h igher median AUC (0.86) than the other three algorithms\, DNN performance was very poor for some species (below 0.6). Balanced Random Forest (AUC me dian 0.82) performed relatively better for all species in comparison to th e other three algorithms. Furthermore\, for some species where the other t hree algorithms performed poorly (AUC less than 70%)\, Balanced-RF outperf orms the others.\n\nThe user experiment results provided us with prelimina ry findings that support the combination of citizen science and machine le arning. According to the findings of the user experiment\, participants wi th a higher number of contributions found real-time feedback to be more us eful in learning about biodiversity and stated that it increased their mot ivation to contribute to the project. Besides that\, as a result of automa tic data validation\, only 10% of observations were flagged for expert ver ification\, resulting in a faster validation process and improved data qua lity by combining human and machine power. \n\n\nWhy it should be consider ed:\n\nData validation and long-term participation have always been two of the most difficult challenges in citizen science and VGI (volunteer geogr aphic information) projects. Various studies have been conducted on biodiv ersity data validation\, focusing primarily on observation images with aut omatic species identification\; however\, not enough attention has been pa id to observation location validation\, particularly automatic location va lidation taking into account species habitat characteristics. Furthermore\ , to the best of our knowledge\, the combination of machine learning and c itizen science for sustaining participation by providing real-time user-ce ntered and machine generated feedback to participants has received\, till now\, little attention and therefore our work is new\, original and comple tely coherent with the vision of community citizen science\, where scienti sts and citizen scientists are supposed to learn from each other. \n\nBibl iography:\n\n\nLand-Zandstra\, Anne M.\, Jeroen L. A. Devilee\, Frans Snik \, Franka Buurmeijer\, and Jos M. van den Broek. 2016. “Citizen Science on a Smartphone: Participants’ Motivations and Learning.” Public Under standing of Science 25 (1): 45–60.\n\nLotfian\, Maryam\, Jens Ingensand \, Olivier Ertz\, Simon Oulevay\, and Thibaud Chassin. July 15-20\, 2019. “Auto-Filtering Validation in Citizen Science Biodiversity Monitoring: A Case Study.” In Proceedings of the 29th ICA Conference. Vol. 2. https:/ /doi.org/10.5194/ica-proc-2-78-2019.\n\nSchade S\, Tsinaraki C.\; Survey r eport: data management in Citizen Science projects\; EUR 27920 EN\; Luxemb ourg (Luxembourg): Publications Office of the European Union\; 2016\; doi: 10.2788/539115 DTSTAMP:20260422T234410Z LOCATION:Room Hall 3A SUMMARY:An approach for real-time validation of the location of biodiversit y observations contributed in a citizen science project - Maryam Lotfian URL:https://talks.osgeo.org/foss4g-2022-academic-track/talk/VMNCM3/ END:VEVENT END:VCALENDAR