Päästesündmuste andmete sidumine avalike andmetega
Date
2017
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Päästeteenistuste töötajad püüavad iga päev tagada, et inimesed elaksid võimalikult turvalises keskkonnas. Nende eesmärk on jõuda päästealase turvalisuse tasemelt võrdsele positsioonile Põhjamaadega. See tähendaks päästesündmuste langust, tõusnud ohutusalast teadlikkust ja suurenenud koostööd partneritega ja kodanikega. Antud lõputöö eesmärk oli uurida seoseid päästesündmuste andmete ja avalike andmete vahel, nagu avalikud üritused ja ilmaandmed, et leida tugevaid korrelatsioone, mis aitaks Päästeametil paremini hinnata tekkivaid riske. Andmete uurimisel rakendas autor korrelatsioonianalüüsi, Sign testi, aegridade ja logistilise regressiooni metoodikaid. Tulemused näitasid, et ilmaandmetel on kõige tugevam seos päästesündmustega „Tulekahjud“, „Abitus seisundis loom/lind“ ja „Nafta-saadustega reostumine“. Samuti tuli välja, et liiklusõnnetusi juhtub rohkem päevadel, kui toimuvad ka peod. Kõige tihedamini juhtub seda nädala lõpus. Lisaks, logistilise regressiooni analüüsi läbiviimisel, tuli välja, et tugevate seostega muutujad avalike ürituste ja ilmastiku andmetest ennustasid päästesündmuste tekkimise tõenäosust üsna nõrgalt.
The workers of the rescue services strive daily to ensure that the people live in a safe environment. Their aim is to achieve the level of safety indicators common to the Nordic countries. It means having fewer accidents, risen preventive awareness and increased co-operation with partners and citizens. The purpose of the thesis was to explore relationships of rescue event data with online data, such as public events and weather indicators, in or-der to find strong correlations that will allow the Rescue Board to estimate rescue event risk changes. The author applied correlational study, sign test method, time series and lo-gistic regression analyses while studying the data. Results has shown that the weather data is in a strongest relationship with “Fires”, “Helpless animal/bird” and “Oil spills” rescue event types. Also, it appeared that road accidents occur more often on days when parties and celebrations occur as well. Most of these days are in the end of the week. Then, while conducting a logistical regression analysis, it appeared that statistically significant variables from public event and weather datasets predicted the probability of occurrence of rescue event considerably poorly.
The workers of the rescue services strive daily to ensure that the people live in a safe environment. Their aim is to achieve the level of safety indicators common to the Nordic countries. It means having fewer accidents, risen preventive awareness and increased co-operation with partners and citizens. The purpose of the thesis was to explore relationships of rescue event data with online data, such as public events and weather indicators, in or-der to find strong correlations that will allow the Rescue Board to estimate rescue event risk changes. The author applied correlational study, sign test method, time series and lo-gistic regression analyses while studying the data. Results has shown that the weather data is in a strongest relationship with “Fires”, “Helpless animal/bird” and “Oil spills” rescue event types. Also, it appeared that road accidents occur more often on days when parties and celebrations occur as well. Most of these days are in the end of the week. Then, while conducting a logistical regression analysis, it appeared that statistically significant variables from public event and weather datasets predicted the probability of occurrence of rescue event considerably poorly.