• Document: NSE Big Data Challenge
  • Size: 3.29 MB
  • Uploaded: 2019-05-17 03:39:49
  • Status: Successfully converted

Some snippets from your converted document:

Time Activity 2.00pm Introduction to NSE Big Data Challenge 2.10pm What is Data Processing? - Sharing on the parameters collected and explanation of NSE Big Data the processed data 2.30pm Demonstration on ModStore Challenge 3.00pm Sharing on the Supercomputing Facilities 3.10pm Hands-on Session for Students 4.00pm End of Workshop Background and Objectives • A nation-wide project launched by President Tony Tan in Jan 2015 • First National-scale deployment of IoT devices designed for ease of use • Involved 176 schools and more than 90,000 students, and in 2 years NSE Big Data Challenge Objectives: To allow students to learn about big data analytics through the use of the NSE data – Processing and filtering of big data – Use of big data tools – Draw meaningful insights from big data – Presentation of analyses in easy-to-understand ways Timeline S/N Date Date 1 Half-day Preparatory 17, 18, 19 Oct Workshop 2 Submission of Entries 9 Dec 2016 3 NSE Big Data Challenge 3rd week of Jan 2017 Finale - Exhibition, Prize Ceremony Prizes 1st prize: • Up to $300 worth of gifts for each member + cash contribution to school student fund 2nd prize: • Up to $200 worth of gifts for each member + cash contribution to school’s student fund 3rd prize: • Up to $100 worth of gifts for each member + cash contribution to school’s student fund Consolation prizes • Up to $50 worth of gifts for each member + cash contribution to school’s student fund Materials to be submitted 1. Written Report*: a) Innovation (25%) • How creative is the use of the data b) Technical Accuracy (25%) • How well is the data processed c) Impact (25%) • How much social, environment, economic value does it create 2. Presentation of Analyses a) Using maps, slides, video, etc. (25%) *6-page report, font size 12, including annexes What Data is Available? How was Data Processed? Variable Explanation aircon_co2 CO2 emissions from aircon aircon_energy Energy consumption of aircon poi_lat Point of interest (POI) latitude poi_lon Point of interest (POI) longitude stairs _climbed Number of stairs climbed travel_co2 CO2 emissions from the transport mode outdoor_time Time spent outdoor am_travel_mode Transport mode in the morning pm_travel_mode Transport mode in the afternoon IHPC Confidential 8 Air-con usage Identified by a temperature threshold + rapid drops/rises in humidity to mark the start & stop times PoI identified if a number of points cluster in a particular area, e.g. 4 points in a 2min span around the same location. School/home PoIs guessed based on time of day, shopping centres, etc. based on any such PoI found in vicinity on Google Maps. Stairs Identified by pressure differentials (just like the air-con usage is identified by humidity differentials). 9 Travel modes Walking trips identified based on speed threshold (e.g. 1m/s), other transport modes also based on speed, and accelerometer patterns. Public transport trips based on number of points along a public transport route using Google Maps. Outdoors time Differentiated by light intensity; bright = outdoors, dark = indoors. 10 What is ModStore? Data science platform to explore, visualize and find insight from data. All on a browser! • No need to download data – access your work anywhere! • No need to write formula • Included a wide range of statistical and visualization tools, e.g. histogram, boxplot, t-test, correlation, heat map • Just drag and drop the appropriate tools to build your own model 11 Process of data investigation Iterative process: 1. What data is available? 2. What are possible problems my team can pose using this data? 3. Explore data 4. Present findings and solutions What data is available in ModStore? Your own school’s raw data • Each time a device uploads raw sensor data, it creates one row in the dataset • E.g. temperature, humidity, noise Your own school’s processed data • Each row in the dataset represents the processed data for each experiment day • E.g. transport mode, distance and duration in the morning and afternoon Your own school’s happy button data (2016 only) • Each time the happy button is pressed, it creates one row in the dataset • 1 = happy, 2

Recently converted files (publicly available):