SlideShare a Scribd company logo
1 of 45
Download to read offline
Urban flooding prediction
Outline
About Transight
Background and Problem
Analytics Story
Conclusion
Q&A
What we do in Transight
Cloud based data
solution deployment
Enterprise scale
data platform with
Pnda
Customized ML
modeling and Apps
Problem
➢ The problem exists not just in developing countries, but also in
developed countries.
➢ Huge private and public property loss during the flood disaster
➢ Life safety of the civilian live in the city
Resource: Smith, B., & Rodriguez, S. (2017). Spatial Analysis of High-Resolution Radar Rainfall
and Citizen-Reported Flash Flood Data in Ultra-Urban New York City. Water, 9(10), 736.
Current solutions
Resources
IoT solutions
Real time flood level monitoring Rain gauge measurement
How to predict the flood severity 3 hrs in advance?
Expectation
Predictive urban flood severity
Low prevention cost for the government
Efficient resource management
A piece of cake?
flood level data
Rainfall data
Regression Model
How hard could it be !?
Recap(1/3): Flood data
timestamp STCD Height
(cm)
2018-06-08 09:05:00 21 16
2018-06-08 09:10:00 21 15
…... …... …...
2018-08-08 09:35:00 170 0
Flood stations
Recap(2/3): Rainfall data
timestamp STCD Hour_Rain
(mm)
2018-08-23 04:00:00 G3659 3.4
2018-08-23 05:00:00 G1120 36.7
……. …... …...
2018-08-23 15:00:00 G3521 15.3
Weather stations (rain
gauge)
Recap(3/3): Stations geo information
STCD Longitude Latitude
21 113.xxx 22.xxx
21 113.xxx 22.xxx
G1166 113.xxx 22.xxx
G3521 113.xxx 22.xxx
Challenges: common issues
No flood VS Flood
● Missing values
● Extremely high values (>300cm)
Challenges of this project
Get the volume of rainfall for each flood station
Flood data collected irregularly
Predict the flood severity of next 3 hours
Challenges and mitigations (1/3)
Calculate the volume of rainfall for each flood station based on distance
Distance-based weighted rainfall
Challenges and mitigations (2/3)
Flood data collected irregularly - frequency based on flood level
Split data by a fixed time window (1 hour)
0m 10m 20m 30m 40m 50m
0
0 7 4
6 22 21 20 21
3 6 3 1 0
0 11 10 10 9
9 9 8 7 5 4
Challenges and mitigations (2/3)
Flood data collected irregularly - frequency based on flood level
Split data by a fixed time window (1 hour)
0m 10m 20m 30m 40m 50m
0
0 7 4
6 22 21 20 21
3 6 3 1 0
0 11 10 10 9
9 9 8 7 5 4
Challenges and mitigations (2/3)
Flood data collected irregularly - frequency based on flood level
Split data by a fixed time window (1 hour)
0m 10m 20m 30m 40m 50m
0 0 0 0 0 0
0 3.5 7 3.5 3.7 4
6 22 21 20 20.5 21
3 6 4.5 3 1 0
0 5.5 11 10 10 9
9 9 8 7 5 4
0m 10m 20m 30m 40m 50m
0
0 7 4
6 22 21 20 21
3 6 3 1 0
0 11 10 10 9
9 9 8 7 5 4
Challenges and mitigations (3/3)
Predict the flood severity of next 3 hours
Traditional multivariate regression, such as SVR, linear regression
Challenges and mitigations (3/3)
Predict the flood severity of next 3 hours
Our Seq2Seq model
m-dimensional vector
Prepare
Data
Train
Models
Evaluate
Models
Deploy
Data preprocessing
Replace outliers: extreme high values
MinMaxScaler
Feature engineering
● Aggregate the hourly flood levels
0m 10m 20m 30m 40m 50m
0 0 0 0 0 0
0 3.5 7 3.5 3.7 4
6 22 21 20 20.5 21
3 6 4.5 3 1 0
0 5.5 11 10 10 9
9 9 8 7 5 4
Aggregated features: count, std, latest, max, mean and
etc.
Calculated rainfall: rain
Station info: flood history of a flood station, i.e. count
Target value: mean
Feature engineering
● Slide windows of t hours
Data enrichment
● Google Maps API
○ Elevations: elevation, west, east, north, south
● Gaode Maps API
○ Geo-location profiling: restaurants, tunnels,
metro stations and etc.
South
North
EastWest
Seq2seq prediction with Encoder-Decoder
Seq2Seq model in a nutshell (1/3)
● Seq2Seq models: recurrent neural networks for sequence to sequence
prediction
● Why?
○ Allow us to process information that has a time or order element to it
○ Preserve information that couldn’t be done via normal neural network
Sentence: [‘I’, ‘am’, ‘a’, ‘student’, ‘.’]
Bitcoin prices: [8191., 8267., 8251., 8261., 8240., 8219., 8212., 8212., 8206., 8155.]
Seq2Seq model in a nutshell (2/3)
● Use cases: neural machine translation, chatbots, auto reply, time series and
more
Luong, et al. "Neural machine translation (seq2seq) tutorial." (2017): 1532-1543.
Seq2Seq in a nutshell (3/3)
● Vanilla RNN, LSTM or GRU
● Single-layer or multiple-layer
LSTM
Back to urban flood prediction
Define the performance metrics
Matching the prediction goal with the performance metrics
● The higher the flood level, the more crucial to predict it correctly
⇒ customized mae: mae_10, mae_20
Define the performance metrics
0 2 0
1.3 3.3 3.3
0 0 0
1 1 1
52.5 52.5 30.3
24 24 3
5.7 7.9 7.9
0 0 2
11 17 17
0 0 0
0 2.8 3.5
0 0 0
0 0 0
57.7 51.9 31.5
24.6 20.8 9
4.7 7.4 8.7
0 0 0
14 12 9
0 2 0
1.3 0.5 0.2
0 0 0
1 1 1
5.2 0.6 1.2
0.6 3.2 6
1 0.5 0.8
0 0 2
3 5 8
y_true y_pred abs error
mae_10 = 3.35 (26.8/8)
mae_20 = 2.16 (10.8/5)
Define the error metrics
● Performance metrics: mae_10, mae_20
● Loss function: mean squared error (MSE)
Build models
LSTM LSTM LSTM
h1 h2
FC FC
LSTM LSTM LSTM
h1 h2
Encoder Decoder
Seq2Seq Baseline
Start
End
X1 X2 Xt
LSTM LSTM LSTM
h1 h2
FC FC
LSTM LSTM LSTM
h1 h2
Q1 Q3
Encoder Decoder
Seq2Seq + weather forecast
Concat Concat
Start
End
X1 X2 Xt
FC
LSTM LSTM LSTM
FC FC
LSTM LSTM LSTM
Concat
FC
Concat
FC
Concat
Encoder Decoder
Seq2Seq + weather + Geoinfo
Q1
Concat
Q3
Concat
Start
End
Q1 Q3
X1 X2 XtS1 S2 S2
Results
● Performance of the above three models
Model rmse mae_10 mae_20
Seq2Seq baseline 0.020 3.783 3.258
Seq2seq + weather
forecast
0.021 3.411 3.005
Seq2Seq + weather
+ geoinfo
0.022 3.322 2.941
Overall performance of the model has reached 60%-80% in
different levels of flood level prediction (20cm to 50cm+)
More about the models
● t =12
○ 12 hrs of history data used to predict the next 3 hours
● Hyperparameters
○ Optimizer: adam
○ epochs=50
Deployment
● Model with weather + geoinfo
● Re-train the model regularly
Wrap up
Results: The model we deployed can predict the flood
level 3 hrs in advance with high reliability.
Action: Government is testing our prediction model
against their experts experience - to list flood severity 3
hrs in advance in the city.
Benefits: Once in production, city will have better flood
prevention and rescue system, save millions per year.
Questions?
Contact us:info@transight.cn
Utrecht, The Netherlands

More Related Content

Similar to Urban flood prediction digital ocean august edition

Ppt gu scs team2_a_metro project
Ppt gu scs team2_a_metro projectPpt gu scs team2_a_metro project
Ppt gu scs team2_a_metro projectMicah Melling
 
IRJET-A Blind Watermarking Algorithm
IRJET-A Blind Watermarking AlgorithmIRJET-A Blind Watermarking Algorithm
IRJET-A Blind Watermarking AlgorithmLalith Kumar
 
A Blind Watermarking Algorithm
A Blind Watermarking AlgorithmA Blind Watermarking Algorithm
A Blind Watermarking AlgorithmIRJET Journal
 
IRJET-A Blind Watermarking Algorithm
IRJET-A Blind Watermarking AlgorithmIRJET-A Blind Watermarking Algorithm
IRJET-A Blind Watermarking AlgorithmIRJET Journal
 
Disaster Debris Detection and Management System using WSN & IoT
Disaster Debris Detection and Management System using WSN & IoTDisaster Debris Detection and Management System using WSN & IoT
Disaster Debris Detection and Management System using WSN & IoTEswar Publications
 
How Deep Learning Could Predict Weather Events
How Deep Learning Could Predict Weather EventsHow Deep Learning Could Predict Weather Events
How Deep Learning Could Predict Weather Eventsinside-BigData.com
 
DSD-INT 2015 - Integrate high preformance 2D Flood Simulation into FEWS - Kev...
DSD-INT 2015 - Integrate high preformance 2D Flood Simulation into FEWS - Kev...DSD-INT 2015 - Integrate high preformance 2D Flood Simulation into FEWS - Kev...
DSD-INT 2015 - Integrate high preformance 2D Flood Simulation into FEWS - Kev...Deltares
 
Full Scale Data Handling in Shipping: A Big Data Solution
Full Scale Data Handling in Shipping: A Big Data SolutionFull Scale Data Handling in Shipping: A Big Data Solution
Full Scale Data Handling in Shipping: A Big Data SolutionLokukaluge Prasad Perera
 
Sataid auto-download
Sataid auto-downloadSataid auto-download
Sataid auto-downloadJMA_447
 
Data Time Travel by Delta Time Machine
Data Time Travel by Delta Time MachineData Time Travel by Delta Time Machine
Data Time Travel by Delta Time MachineDatabricks
 
Design and implementation of GPS Tracker
Design and implementation of GPS TrackerDesign and implementation of GPS Tracker
Design and implementation of GPS TrackerVignesh Kannan
 
Mid Term Project Presentation- GSM Based Flood Notification System
Mid Term Project Presentation- GSM Based Flood Notification SystemMid Term Project Presentation- GSM Based Flood Notification System
Mid Term Project Presentation- GSM Based Flood Notification SystemAnshul Joshi
 
Global grid of master events for waveform cross correlation: design and testing
Global grid of master events for waveform cross correlation: design and testingGlobal grid of master events for waveform cross correlation: design and testing
Global grid of master events for waveform cross correlation: design and testingIvan Kitov
 
Drinking Water Networks: Challenges and opportunites
Drinking Water Networks: Challenges and opportunitesDrinking Water Networks: Challenges and opportunites
Drinking Water Networks: Challenges and opportunitesPantelis Sopasakis
 
BWC Supercomputing 2008 Presentation
BWC Supercomputing 2008 PresentationBWC Supercomputing 2008 Presentation
BWC Supercomputing 2008 Presentationlilyco
 
Automatic and interactive spot check of the IDC bulletins (REB, SEL3) and XSE...
Automatic and interactive spot check of the IDC bulletins (REB, SEL3) and XSE...Automatic and interactive spot check of the IDC bulletins (REB, SEL3) and XSE...
Automatic and interactive spot check of the IDC bulletins (REB, SEL3) and XSE...ivanokitov
 
Prof. Sameh Saad - A Digital Model for 3D Characterization of Groundwater Qua...
Prof. Sameh Saad - A Digital Model for 3D Characterization of Groundwater Qua...Prof. Sameh Saad - A Digital Model for 3D Characterization of Groundwater Qua...
Prof. Sameh Saad - A Digital Model for 3D Characterization of Groundwater Qua...Hudhaib Al-Allatti
 
FV_IGARSS11.ppt
FV_IGARSS11.pptFV_IGARSS11.ppt
FV_IGARSS11.pptgrssieee
 

Similar to Urban flood prediction digital ocean august edition (20)

Ppt gu scs team2_a_metro project
Ppt gu scs team2_a_metro projectPpt gu scs team2_a_metro project
Ppt gu scs team2_a_metro project
 
IRJET-A Blind Watermarking Algorithm
IRJET-A Blind Watermarking AlgorithmIRJET-A Blind Watermarking Algorithm
IRJET-A Blind Watermarking Algorithm
 
A Blind Watermarking Algorithm
A Blind Watermarking AlgorithmA Blind Watermarking Algorithm
A Blind Watermarking Algorithm
 
IRJET-A Blind Watermarking Algorithm
IRJET-A Blind Watermarking AlgorithmIRJET-A Blind Watermarking Algorithm
IRJET-A Blind Watermarking Algorithm
 
Disaster Debris Detection and Management System using WSN & IoT
Disaster Debris Detection and Management System using WSN & IoTDisaster Debris Detection and Management System using WSN & IoT
Disaster Debris Detection and Management System using WSN & IoT
 
How Deep Learning Could Predict Weather Events
How Deep Learning Could Predict Weather EventsHow Deep Learning Could Predict Weather Events
How Deep Learning Could Predict Weather Events
 
DSD-INT 2015 - Integrate high preformance 2D Flood Simulation into FEWS - Kev...
DSD-INT 2015 - Integrate high preformance 2D Flood Simulation into FEWS - Kev...DSD-INT 2015 - Integrate high preformance 2D Flood Simulation into FEWS - Kev...
DSD-INT 2015 - Integrate high preformance 2D Flood Simulation into FEWS - Kev...
 
Full Scale Data Handling in Shipping: A Big Data Solution
Full Scale Data Handling in Shipping: A Big Data SolutionFull Scale Data Handling in Shipping: A Big Data Solution
Full Scale Data Handling in Shipping: A Big Data Solution
 
Sataid auto-download
Sataid auto-downloadSataid auto-download
Sataid auto-download
 
Data Time Travel by Delta Time Machine
Data Time Travel by Delta Time MachineData Time Travel by Delta Time Machine
Data Time Travel by Delta Time Machine
 
Design and implementation of GPS Tracker
Design and implementation of GPS TrackerDesign and implementation of GPS Tracker
Design and implementation of GPS Tracker
 
Mid Term Project Presentation- GSM Based Flood Notification System
Mid Term Project Presentation- GSM Based Flood Notification SystemMid Term Project Presentation- GSM Based Flood Notification System
Mid Term Project Presentation- GSM Based Flood Notification System
 
Time series Forecasting
Time series ForecastingTime series Forecasting
Time series Forecasting
 
Global grid of master events for waveform cross correlation: design and testing
Global grid of master events for waveform cross correlation: design and testingGlobal grid of master events for waveform cross correlation: design and testing
Global grid of master events for waveform cross correlation: design and testing
 
Drinking Water Networks: Challenges and opportunites
Drinking Water Networks: Challenges and opportunitesDrinking Water Networks: Challenges and opportunites
Drinking Water Networks: Challenges and opportunites
 
Presentation
PresentationPresentation
Presentation
 
BWC Supercomputing 2008 Presentation
BWC Supercomputing 2008 PresentationBWC Supercomputing 2008 Presentation
BWC Supercomputing 2008 Presentation
 
Automatic and interactive spot check of the IDC bulletins (REB, SEL3) and XSE...
Automatic and interactive spot check of the IDC bulletins (REB, SEL3) and XSE...Automatic and interactive spot check of the IDC bulletins (REB, SEL3) and XSE...
Automatic and interactive spot check of the IDC bulletins (REB, SEL3) and XSE...
 
Prof. Sameh Saad - A Digital Model for 3D Characterization of Groundwater Qua...
Prof. Sameh Saad - A Digital Model for 3D Characterization of Groundwater Qua...Prof. Sameh Saad - A Digital Model for 3D Characterization of Groundwater Qua...
Prof. Sameh Saad - A Digital Model for 3D Characterization of Groundwater Qua...
 
FV_IGARSS11.ppt
FV_IGARSS11.pptFV_IGARSS11.ppt
FV_IGARSS11.ppt
 

Recently uploaded

Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupFlorian Wilhelm
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitecturePixlogix Infotech
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):comworks
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationSlibray Presentation
 
Benefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other FrameworksBenefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other FrameworksSoftradix Technologies
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticscarlostorres15106
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphNeo4j
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationRidwan Fadjar
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubKalema Edgar
 
Pigging Solutions Piggable Sweeping Elbows
Pigging Solutions Piggable Sweeping ElbowsPigging Solutions Piggable Sweeping Elbows
Pigging Solutions Piggable Sweeping ElbowsPigging Solutions
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsAndrey Dotsenko
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Patryk Bandurski
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationSafe Software
 

Recently uploaded (20)

E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptxE-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project Setup
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC Architecture
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck Presentation
 
Benefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other FrameworksBenefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other Frameworks
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food Manufacturing
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 Presentation
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding Club
 
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort ServiceHot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
 
Pigging Solutions Piggable Sweeping Elbows
Pigging Solutions Piggable Sweeping ElbowsPigging Solutions Piggable Sweeping Elbows
Pigging Solutions Piggable Sweeping Elbows
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
 
The transition to renewables in India.pdf
The transition to renewables in India.pdfThe transition to renewables in India.pdf
The transition to renewables in India.pdf
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
 

Urban flood prediction digital ocean august edition

  • 2. Outline About Transight Background and Problem Analytics Story Conclusion Q&A
  • 3. What we do in Transight Cloud based data solution deployment Enterprise scale data platform with Pnda Customized ML modeling and Apps
  • 4.
  • 5. Problem ➢ The problem exists not just in developing countries, but also in developed countries. ➢ Huge private and public property loss during the flood disaster ➢ Life safety of the civilian live in the city Resource: Smith, B., & Rodriguez, S. (2017). Spatial Analysis of High-Resolution Radar Rainfall and Citizen-Reported Flash Flood Data in Ultra-Urban New York City. Water, 9(10), 736.
  • 7. IoT solutions Real time flood level monitoring Rain gauge measurement
  • 8. How to predict the flood severity 3 hrs in advance?
  • 9. Expectation Predictive urban flood severity Low prevention cost for the government Efficient resource management
  • 10. A piece of cake? flood level data Rainfall data Regression Model
  • 11. How hard could it be !?
  • 12. Recap(1/3): Flood data timestamp STCD Height (cm) 2018-06-08 09:05:00 21 16 2018-06-08 09:10:00 21 15 …... …... …... 2018-08-08 09:35:00 170 0 Flood stations
  • 13. Recap(2/3): Rainfall data timestamp STCD Hour_Rain (mm) 2018-08-23 04:00:00 G3659 3.4 2018-08-23 05:00:00 G1120 36.7 ……. …... …... 2018-08-23 15:00:00 G3521 15.3 Weather stations (rain gauge)
  • 14. Recap(3/3): Stations geo information STCD Longitude Latitude 21 113.xxx 22.xxx 21 113.xxx 22.xxx G1166 113.xxx 22.xxx G3521 113.xxx 22.xxx
  • 15. Challenges: common issues No flood VS Flood ● Missing values ● Extremely high values (>300cm)
  • 16. Challenges of this project Get the volume of rainfall for each flood station Flood data collected irregularly Predict the flood severity of next 3 hours
  • 17. Challenges and mitigations (1/3) Calculate the volume of rainfall for each flood station based on distance Distance-based weighted rainfall
  • 18. Challenges and mitigations (2/3) Flood data collected irregularly - frequency based on flood level Split data by a fixed time window (1 hour) 0m 10m 20m 30m 40m 50m 0 0 7 4 6 22 21 20 21 3 6 3 1 0 0 11 10 10 9 9 9 8 7 5 4
  • 19. Challenges and mitigations (2/3) Flood data collected irregularly - frequency based on flood level Split data by a fixed time window (1 hour) 0m 10m 20m 30m 40m 50m 0 0 7 4 6 22 21 20 21 3 6 3 1 0 0 11 10 10 9 9 9 8 7 5 4
  • 20. Challenges and mitigations (2/3) Flood data collected irregularly - frequency based on flood level Split data by a fixed time window (1 hour) 0m 10m 20m 30m 40m 50m 0 0 0 0 0 0 0 3.5 7 3.5 3.7 4 6 22 21 20 20.5 21 3 6 4.5 3 1 0 0 5.5 11 10 10 9 9 9 8 7 5 4 0m 10m 20m 30m 40m 50m 0 0 7 4 6 22 21 20 21 3 6 3 1 0 0 11 10 10 9 9 9 8 7 5 4
  • 21. Challenges and mitigations (3/3) Predict the flood severity of next 3 hours Traditional multivariate regression, such as SVR, linear regression
  • 22. Challenges and mitigations (3/3) Predict the flood severity of next 3 hours Our Seq2Seq model m-dimensional vector
  • 24. Data preprocessing Replace outliers: extreme high values MinMaxScaler
  • 25. Feature engineering ● Aggregate the hourly flood levels 0m 10m 20m 30m 40m 50m 0 0 0 0 0 0 0 3.5 7 3.5 3.7 4 6 22 21 20 20.5 21 3 6 4.5 3 1 0 0 5.5 11 10 10 9 9 9 8 7 5 4 Aggregated features: count, std, latest, max, mean and etc. Calculated rainfall: rain Station info: flood history of a flood station, i.e. count Target value: mean
  • 26. Feature engineering ● Slide windows of t hours
  • 27. Data enrichment ● Google Maps API ○ Elevations: elevation, west, east, north, south ● Gaode Maps API ○ Geo-location profiling: restaurants, tunnels, metro stations and etc. South North EastWest
  • 28. Seq2seq prediction with Encoder-Decoder
  • 29. Seq2Seq model in a nutshell (1/3) ● Seq2Seq models: recurrent neural networks for sequence to sequence prediction ● Why? ○ Allow us to process information that has a time or order element to it ○ Preserve information that couldn’t be done via normal neural network Sentence: [‘I’, ‘am’, ‘a’, ‘student’, ‘.’] Bitcoin prices: [8191., 8267., 8251., 8261., 8240., 8219., 8212., 8212., 8206., 8155.]
  • 30. Seq2Seq model in a nutshell (2/3) ● Use cases: neural machine translation, chatbots, auto reply, time series and more Luong, et al. "Neural machine translation (seq2seq) tutorial." (2017): 1532-1543.
  • 31. Seq2Seq in a nutshell (3/3) ● Vanilla RNN, LSTM or GRU ● Single-layer or multiple-layer LSTM
  • 32. Back to urban flood prediction
  • 33. Define the performance metrics Matching the prediction goal with the performance metrics ● The higher the flood level, the more crucial to predict it correctly ⇒ customized mae: mae_10, mae_20
  • 34. Define the performance metrics 0 2 0 1.3 3.3 3.3 0 0 0 1 1 1 52.5 52.5 30.3 24 24 3 5.7 7.9 7.9 0 0 2 11 17 17 0 0 0 0 2.8 3.5 0 0 0 0 0 0 57.7 51.9 31.5 24.6 20.8 9 4.7 7.4 8.7 0 0 0 14 12 9 0 2 0 1.3 0.5 0.2 0 0 0 1 1 1 5.2 0.6 1.2 0.6 3.2 6 1 0.5 0.8 0 0 2 3 5 8 y_true y_pred abs error mae_10 = 3.35 (26.8/8) mae_20 = 2.16 (10.8/5)
  • 35. Define the error metrics ● Performance metrics: mae_10, mae_20 ● Loss function: mean squared error (MSE)
  • 37. LSTM LSTM LSTM h1 h2 FC FC LSTM LSTM LSTM h1 h2 Encoder Decoder Seq2Seq Baseline Start End X1 X2 Xt
  • 38. LSTM LSTM LSTM h1 h2 FC FC LSTM LSTM LSTM h1 h2 Q1 Q3 Encoder Decoder Seq2Seq + weather forecast Concat Concat Start End X1 X2 Xt
  • 39. FC LSTM LSTM LSTM FC FC LSTM LSTM LSTM Concat FC Concat FC Concat Encoder Decoder Seq2Seq + weather + Geoinfo Q1 Concat Q3 Concat Start End Q1 Q3 X1 X2 XtS1 S2 S2
  • 40. Results ● Performance of the above three models Model rmse mae_10 mae_20 Seq2Seq baseline 0.020 3.783 3.258 Seq2seq + weather forecast 0.021 3.411 3.005 Seq2Seq + weather + geoinfo 0.022 3.322 2.941 Overall performance of the model has reached 60%-80% in different levels of flood level prediction (20cm to 50cm+)
  • 41. More about the models ● t =12 ○ 12 hrs of history data used to predict the next 3 hours ● Hyperparameters ○ Optimizer: adam ○ epochs=50
  • 42. Deployment ● Model with weather + geoinfo ● Re-train the model regularly
  • 43. Wrap up Results: The model we deployed can predict the flood level 3 hrs in advance with high reliability. Action: Government is testing our prediction model against their experts experience - to list flood severity 3 hrs in advance in the city. Benefits: Once in production, city will have better flood prevention and rescue system, save millions per year.