Da Data Science and It Its Applications in Smart Cities Mohadeseh Ganji, Senior Data Scientist, Research fellow 15 Nov 2017
About me Currently, I am a post-doctoral research fellow at University of Melbourne Mohadeseh.ganji@unimelb.edu.au PhD in Machine learning from University of Melbourne Researcher at CSIRO, Data61. 1+1 Taught Machine Learning at Melbourne Business School Industry experience in finance and ICT sectors Pixabay free illustrations
Data Is Everywhere ! 4.6 billion 30+ billion RFID camera phones 3.5 billions of tags world wide search query ? TBs of every day data every day 12+ TBs of tweet data every day 100s of millions 100 hours of of GPS enabled 25+ TBs of video uploads devices sold annually log data every per minutes day 6.5 PB of user data 3.8+ billion people on 50 TB/ day 200 million smart the Web meters
World Economic Forum
Data Science Connected Devices Lab
Data Science Connected Devices Lab
Data Science: The Science of Translating Data to Wisdom
Knowledge Comes From Data! Chris Anderson
Parts of a Data Science Project • Collection: getting the data • Engineering: storage and computational resources across full lifecycle • Governance: overall management of data across full lifecycle • Wrangling: data preprocessing, cleaning • Analysis: discovery (learning, visualisation, etc. ) • Presentation: arguing the case that the results are significant and useful • Operationalisation: putting the results to work, so as to gain benefits or value We call this the Standard Value Chain.
Data Science: A Multidisciplinary Science Drew Conway’s Venn diagram
Main Data Science Algorithms Descriptive Predictive
Descriptive Analysis Data science algorithms that describe data and provide insight A Collection Descriptive Patterns / of Data Algorithms Insights
Predictive Analysis Data science algorithms that make predictions
Often You Need a Combination Images Video Social Geospatial Media Data Descriptive and Patterns / Insight / Predictive Recommendations Algorithms / Actions Text Meters Data Sensors Data GPS Data
What is Big Data? Here is from Wikipedia: • Big data is the term for a collection of data sets so large and complex that it becomes difficult to process using on-hand database management tools or traditional data processing applications. • The challenges include capture, curation, storage, search, sharing, transfer, analysis, and visualization. • Enabled by the cloud: affordability, extensibility, agility 15
The four Vs of the Big Data IBM inforgraphic
Where Are The Boundaries in Data Science? Machine Learning Artificial Data Intelligence Science Deep Learning
Application of Data Science in Smart Cities Free vector graphics on Pixabay
Smart Transportation Implementing smart traffic light and signals Better utilization of parking designed by the Traffic21 space Monitoring of parking spaces project in Pittsburgh, availability in the city. Pennsylvania, USA reduced traffic jams and waiting times and resulted in reduced emissions by over 20% Optimize traffic flow using traffic signals, the number of vehicles and pedestrians; Recognize traffic patterns by analysing the data Reduce road congestions by predicting traffic conditions and adjusting traffic controls, alternative roads or informing commuters
Smart Transportation Communicate to drivers using on- vehicle devices to inform them Better utilization of parking about traffic situation or to take Traffic accident prediction (crash space action to alleviate the problem. frequency and crash severity) Monitoring of parking spaces availability in the city. Pedestrian behaviour analysis In different traffic situation, weather, etc. Crossing behaviour Optimize traffic flow using Transport infrastructure traffic signals, the number of maintenance analysis vehicles and pedestrians; Recognize traffic patterns by analysing the data Travel time prediction using Bluetooth data and information on traffic situation Reduce road congestions by predicting traffic conditions and adjusting traffic controls, alternative roads or informing commuters Better public transportation planning using the tap-in tap-out data
Smart Disaster Management Predict future environmental changes or natural disasters Using historical data, spatial temporal data, Natural hazard management using social network satellite information etc. data generated by citizens and first responders, spatial temporal data, satellite images of the affected areas, flood maps generated by drones. Road network resilience, traffic demand prediction during disaster ( e.g. flash floods and bush fires) Manage information flow in disaster Prioritize urgent sites and situations. Responders social media mining and information want up-to-date intelligence about survivors' dissemination during disasters. locations and available resources. Data analytics shows these kinds of details. Actual mapping applications are emerging that can prioritize which zones need attention first, all based on what users Track resources. Data analytics can help map post during a crisis. the locations of critical resources like ambulances and medical facilities. Monitoring , analyzing and identifying the risks vibrations and material conditions data in buildings, bridges and historical monuments along with data from from weather forecasts, geologic surveys, maintenance reports, video feeds to detect unusual patterns, identify red-light situations and create a clearer picture of risk
Smart Energy Smart grid Total potential value • Supplier consumer behaviour, generated in the United • Minute or second level data States from a fully from sensors and meters on production, transmission, deployed smart grid distribution systems and reaching as high as $130 consumer access points. billion annually by 2019. • Two way communication between producers and consumer McKinsey on Smart Grid
Smart Energy Accident prevention Smart monitoring the infrastructure and analyse the data Smart grid • Supplier consumer behaviour, Network Reliability prevent power outages, • Minute or second level data interruptions and quality issues from sensors and meters on production, transmission, distribution systems and Load Modeling: consumer access points. Understanding the behaviour • Two way communication of the individual and system between producers and in different situations consumer Smart Buildings: Optimize building electricity usage Demand forecasting with motion sensor lights capacity planning, demand- which can dim or shut off when response modelling and a room is empty; Alert when power distribution there is a leaking pipe using smart meters; monitor energy Smart Pricing based on use of an electric meter and demand and supply data alert when it reaches a specific threshold
Smart Healthcare Analyzing disease patterns : Patient profiling Smart analyzing disease patterns, trends gathering, analyzing and utilizing and spread patterns for prevention of patient information and make strategic decisions Better statistical tools and algorithms to improve clinical trial design Identify at risk patients: Based on the patient profile, Monitor, analyze and flag Healthcare to potential health issues Patient stay and treatment outcome identify who would benefit from Analytics/Medical Analytics prediction to study patient characteristics proactive care or lifestyle change Market is expected to reach and the cost and outcomes of treatments around 18.7 Billion USD by Smart health monitoring devices 2020 at a CAGR of 26.5% Smart monitoring of blood sugar, from 2015 to 2020. blood pressure, sleep patterns for Resource scheduling based on demand accurate and timely responses to prediction, health issues www.marketwatch.com Medical staff, equipment, ambulances,… Health Economics: performance- Medical decision support systems : based pricing plans based on real- Diagnosing and treatment Personalized medicine : understanding world patient outcomes data to arrive at genetic variation and individual treatment fair economic compensation response
Smart Healthcare Analyzing disease patterns : Patient profiling Smart analyzing disease patterns, trends gathering, analyzing and utilizing and spread patterns for prevention of patient information and make strategic decisions Better statistical tools and algorithms to improve clinical trial design Identify at risk patients: Based on the patient profile, Monitor, analyze and flag to potential health issues Patient stay and treatment outcome identify who would benefit from prediction to study patient characteristics proactive care or lifestyle change and the cost and outcomes of treatments Smart health monitoring devices Smart monitoring of blood sugar, blood pressure, sleep patterns for Resource scheduling based on demand accurate and timely responses to prediction, health issues Medical staff, equipment, ambulances,… Health Economics: performance- Medical decision support systems : based pricing plans based on real- Diagnosing and treatment Personalized medicine : understanding world patient outcomes data to arrive at genetic variation and individual treatment fair economic compensation response
Recommend
More recommend
