Civic CrowdAnalytics

Transcript

1. Machine Learning para análisis de datos cívicos Jorge Saldivar Tanja Aitamurto Kaiping Chen Ahmed Cherif Luis Santana
3. Voto
4. Crowdsourcing ideas y soluciones para leyes y reglamentaciones
5. Crowdsourced urban planning strategies
6. Hillary Gitelman, Urbanista, Ciudad de Palo Alto, CA, USA
7. Pilas y pilas de datos desestructurados Palo Alto will have to increase density, not just because ABAG requires more housing units, but because when we add more housing units, at least some of those will be occupied by people employed here, thereby reducing auto trips. Palo Alto must also have additional low-income housing units, to enable at least some of the low wage earners to live nearby (e.g., especially caregivers for seniors). Programs T-1, T-2, and T-3 are all useful approaches. Single family homeowners should be allowed and encouraged to add small in-law units (and be relieved of the current parking regulations for second units). Single-level parking lots should have housing above them (e.g,, behind the retail stores on California Avenue). Consider some commercial zoning within neighborhoods so that markets, coffee shops, hardware stores can be walked to by residents. Current commercial activites are limited to "centers" / University/California/Middlefield/San Antonio and lots of people can't walk to them, so everyone drives, causing more cars on the road and congestion in the parking areas. Parking should be unbundled - i.e. you pay if you want to use it. Lots and lots of young people and seniors would love to live in housing where they're not implicitly paying for parking they don't use and lots of them don't need or want a car, particularly if they're living right near the Caltrain. We should also experiment with housing that doesn't have any parking at all in areas where we have parking permits. Just never give those people permits. Part of the problem with mixed use development is that the FAR for the housing portion is really small. We should be encouraging ground floor retail + 3 floors of housing or retail + commercial + 2 floors of housing. But the current restrictions don't allow you to build that much housing, which exacerbates the jobs housing imbalance. We have never actually instituted density minimums but really need to. We should no long be building single homes or gigantic condos within half a mile of train station. We should make this actually happen. All of the ideas below, for enhancing flow, could be supported by having the Comp Plan delineate how bad the congestion can be at each intersection (LOS) and not allowing development to push it higher than that. The Comp Plan could state that the LOS at a specific intersection would have to IMPROVE in order for more development to be allowed. Delineating the LOS required might be the only remedy the city has to the gridlock- enabling loophole in CEQA . (The one that says CEQA cannot be used in relationship to gridlock isn't it Orwellian that this CEQA protection has been eliminated?) To build, developers would have to consider HOW THEY can take responsibility for the impact of greater density, rather than shifting the burden to the existing residents or the public in general.
8. Cuello de botella en el canal de participación Input de los ciudadanos Ley/Norma/Regulación Falta de herramientas de análisis: la paradoja de la participación
10. Categorización Análisis de sentimiento; Búsqueda de conceptos relacionados Powered by Civic CROWDANALYTICS Aplicación web que utilizar procesamiento de lenguaje natural y machine learning para el análisis de datos cívicos
11. Categorización
12. Categorización en niveles
13. Sentiment analysis
14. CASO DE ESTUDIO • Confección del plan maestro de urbanización de la ciudad de Palo Alto, CA. Transporte urbano en específico
15. CASO DE ESTUDIO • Confección del plan maestro de urbanización de la ciudad de Palo Alto, CA. Transporte urbano en específico • Comité de representantes ciudadanos (CAC)
16. CASO DE ESTUDIO • Confección del plan maestro de urbanización de la ciudad de Palo Alto, CA. Transporte urbano en específico • Comité de representantes ciudadanos (CAC) • Pregunta de investigación: ¿hasta que punto el input de los ciudadanos fue tenido cuenta en el borrador del plan maestro de transporte urbano?
17. CASO DE ESTUDIO • Confección del plan maestro de urbanización de la ciudad de Palo Alto, CA. Transporte urbano en específico • Comité de representantes ciudadanos (CAC) • Pregunta de investigación: ¿hasta que punto el input de los ciudadanos fue tenido cuenta en el borrador del plan maestro de transporte urbano? • Dataset: 184 ideas de ciudadanos + 132 propuestas de los representantes + borrador
18. RESULTADOS PRELIMINARES Cuanto mayor es el volumen (categorización) y más fuerte el tono (análisis de sentimientos) en las propuestas de los ciudadanos más posibilidades tienen de formar parte del borrador
19. Categoría Sentiment Score Promedio % de Sugerencias Neutrales Proporción de input ciudadano Proporción de ideas en el borrador Big picture 0.114 34.7% 27.07% 38.94% Non-motor powered vehicles 0.195 57.14% 15.47% 4.42% Private transit 0.287 44.68% 25.97% 27.43% Public transit 0.206 53.49% 23.76% 26.55% Special needs 0.093 42.86% 7.73% 2.65%
20. RESULTADOS PRELIMINARES El contenido del borrador se asemeja más a lo propuesto por los ciudadanos que a lo sugerido por el comité de representantes (CAC)
21. Términos más frecuentes Input de los ciudadanos Parking (87), cars (53), drive (43) Input de la CAC Downtown (80), traffic congestion(55), sustainable transportation (49) Borrador Parking (37), improvement (28), vehicles (12) TÉRMINOS MÁS FRECUENTES EN LOS DATASETS
22. RESULTADOS PRELIMINARES • Precisión aceptable. Se logró un ratio de precisión (accuracy) de alrededor de 80% en la categorización
23. RESULTADOS PRELIMINARES • Precisión aceptable. Se logró un ratio de precisión (accuracy) de alrededor de 80% en la categorización • Coding es trabajoso. Preparar el dataset para el entrenamiento conlleva mucho tiempo y recursos
24. De pilas y pilas de datos desestructurados… Palo Alto must also have additional low-income housing units, to enable at least some of the low wage earners to live nearby (e.g., especially caregivers for seniors). Programs T-1, T-2, and T-3 are all useful approaches. Single family homeowners should be allowed and encouraged to add small in-law units (and be relieved of the current parking regulations for second units). Single-level parking lots should have housing above them (e.g,, behind the retail stores on California Avenue). Consider some commercial zoning within neighborhoods so that markets, coffee shops, hardware stores can be walked to by residents. Current commercial activites are limited to "centers" / University/California/Middlefield/San Antonio and lots of people can't walk to them, so everyone drives, causing more cars on the road and congestion in the parking areas. Parking should be unbundled - i.e. you pay if you want to use it. Lots and lots of young people and seniors would love to live in housing where they're not implicitly paying for parking they don't use and lots of them don't need or want a car, particularly if they're living right near the Caltrain. We should also experiment with housing that doesn't have any parking at all in areas where we have parking permits. Just never give those people permits. Part of the problem with mixed use development is that the FAR for the housing portion is really small. We should be encouraging ground floor retail + 3 floors of housing or retail + commercial + 2 floors of housing. But the current restrictions don't allow you to build that much housing, which exacerbates the jobs housing imbalance. We have never actually instituted density minimums but really need to. We should no long be building single homes or gigantic condos within half a mile of train station. We should make this actually happen. All of the ideas below, for enhancing flow, could be supported by having the Comp Plan delineate how bad the congestion can be at each intersection (LOS) and not allowing development to push it higher than that. The Comp Plan could state that the LOS at a specific intersection would have to IMPROVE in order for more development to be allowed. Delineating the LOS required might be the only remedy the city has to the gridlock-enabling loophole in CEQA . (The one that says CEQA cannot be used in relationship to gridlock isn't it Orwellian that this CEQA protection has been eliminated?) To build, developers would have to consider HOW THEY can take responsibility for the impact of greater density, rather than shifting the burden to the existing residents or the public in general.
25. … a resultados estructurados
26. proceso online de reforma constitucional en Chile 100.000 participantes 8.500 ideas
27. Tanja Aitamurto Kaiping Chen Ahmed Cherif Jorge Saldivar Luis Santana Gracias!