IS428 AY2019-20T1 Assign Lee Hui Xin Anne

From Visual Analytics for Business Intelligence
Revision as of 13:02, 12 October 2019 by Anne.lee.2017 (talk | contribs) (→‎Dispatch Priority Dashboard)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Problem & Motivation


St. Himark has been hit by an earthquake, leaving officials scrambling to determine the extent of the damage and dispatch limited resources to the areas in most need. They quickly receive seismic readings and use those for an initial deployment but realize they need more information to make sure they have a realistic understanding of the true conditions throughout the city.

In a prescient move of community engagement, the city had released a new damage reporting mobile application shortly before the earthquake. This app allows citizens to provide more timely information to the city to help them understand damage and prioritize their response. In this mini-challenge, use app responses in conjunction with shake maps of the earthquake strength to identify areas of concern and advise emergency planners. Note: the shake maps are from April 6 and April 8 respectively.

With emergency services stretched thin, officials are relying on citizens to provide them with much needed information about the effects of the quake to help focus recovery efforts.

By combining seismic readings of the quake, responses from the app, and background knowledge of the city, help the city triage their efforts for rescue and recovery.

Task Questions:

  1. Emergency responders will base their initial response on the earthquake shake map. Use visual analytics to determine how their response should change based on damage reports from citizens on the ground. How would you prioritize neighborhoods for response? Which parts of the city are hardest hit?
  2. Use visual analytics to show uncertainty in the data. Compare the reliability of neighborhood reports. Which neighborhoods are providing reliable reports? Provide a rationale for your response. Limit your response to 1000 words and 10 images.
  3. How do conditions change over time? How does uncertainty in change over time? Describe the key changes you see. Limit your response to 500 words and 8 images.

Information Gathering

Before the start of analysis, it is essential to have a better understanding about the different areas in St. Himark. Looking through the layout and description of St. Himark several key information was identified to be helpful and critical to the investigation.

Danger Facilities

Using St. Himark City description, some facilities were identified to be threat to the people living in the vicinity. Location 3, SAFE TOWN, a neighborhood where a Nuclear Power Plant resides. Nuclear Power Plants are vulnerable to earthquake and may potentially impact people over a large area causing damage by fires, explosion and release of radioactive materials. In addition to the potential harm it may cause, Residents in Location 3, SAFE TOWN, consists of mostly older-single family who lives close to the Nuclear Power Plant. In case of Earthquake, the residents might not be able to evacuate quickly and might require addition assistance. If no efficient evacuation is provided, SAFE TOWN might face the highest casualty.

Hospital Facilities

Subsequently emergency first aid facilities such as hospitals was identified:

  • Medical Surgical Care located at PALACE HILL
  • OLD TOWN Hospital known for its expertise in diabetes and endocrinology, digestive health, oncology, and orthopedics.
  • DOWNTOWN Himark has Children hospital and has the only Trauma Hospital in the area.
  • BROADVIEW has a Community Hospital which offers a full spectrum of medical care.
  • TERRAPIN SPRINGS has a full-service, non-profit community hospital, offering comprehensive medical, surgical and therapeutic services.
  • SOUTHON able to provide critical care, neurology, gastrointestinal, and orthopedic care.

Using the hospital locations, DOWNTOWN is identified as the area where most causality should be allocated in times of emergency.

Potential Obstructions

Lastly, construction areas was identified to have a significant impact on the efficiency of the evacuation. With Location 6, DOWNTOWN, being the priority place to allocate casualty, the ongoing Road construction in Location 6, DOWNTOWN might cause a delay in the evacuation and unable to provide citizens in need of medical aid. It is essential to identify an alternative road to ensure medical aid is provided in time. Also, with Friday Bridge and Magritte and Jade Bridges being repaired, in anticipation of delays of manpower and medical aid, emergency planner should be advised to request for an air-transportation.

Data Analysis & Transformation Process

The data-sets usually requires cleaning and transformation before it can be sensibly analyzed. It is important to have a proper understanding of the Data, knowing whether the numbers are Categorical or Continuous data and to ensure that it is in the right format. This section elaborates on the data-set analysis and transformation process before it was imported for Interactive Visualization Analysis.

The Data was collected using the Rumble App which crowd sources damage reporting, by collecting information about reporting time, location and different area’s shake intensity with 0 being the lowest and 10 being the highest.

The following section illustrates the issues faced in the data analysis phase leading to a need to transform the data into specified format.


franeless
Issue Faced 1

Issue: The report data provided the shake intensity of various facilities in their specific rows, making analysis by various facilities difficult.

Solution: Data cleaning was performed to pivot the various facilities into a single column called Area using Tableau Prep. By doing so, it reduce the need to perform binning on Tableau during the analysis and it allows Shake Intensity and Damage to be filtered and visualized according to the various areas.

To solve this issue:

  1. Load Tableau Prep and connect data source
  2. Add a Clean up step to review data
  3. Add the Pivot Step and select the relevant variables to be pivoted such as Sewer and water, Power, Road and Bridges, Medical and Buildings
Issue2.png
Issue Faced 2

Issue: To find out which areas has the highest intensity ratings. Filtered by Area, I was only able to view the data by in each instances. Even with St Himark’s map image and location, without the zone coordinated of each areas, the layout could not be plotted. It may posed as a Visualization problem as it is essential for emergency advisers to have an overview of St. Himark rather than having to look through the data instances.

Solution 1

To resolve this issue, the geometrical data of St Himark Neighborhood was made available in a Shape file which can be easily uploaded to tableau through the following steps:

  1. Make sure the shape file is zipped
  2. Open Tableau and connect shape file from the start page
  3. Connecting the geometrical data-set's Location ID to the data-set's Location ID
  4. Now a Geometrical Map can be plot!
Solution 2

An alternative solution would be to identify the coordinates for each zones. This is done using an online tool that allow analysts to get coordinates for custom polygons on a map image. By plotting the neighborhoods on a custom polygon map, it would generate the coordinates of the zone. This will allow us to plot the location of each neighborhood on the image map. The following details the process of generating the individual zone coordinates:

  1. Download the online tool from the following link: https://github.com/bryantbhowell/tableau-map-pack/blob/master/draw_tableau_polygons_on_background_image.html
  2. Open the tool and select the St Himark Image to be plotted. The coordinates was plot based on the St Himark City map provided in the city description. During the plotting process, the areas were based on rough estimation.
  3. After the polygons have been plotted, the results can be exported into a .csv file format for analysis in Tableau. Once the St Himark’s coordinate has been imported into Tableau, the locations of each coordinates will be plotted onto the map. After the completion of coordinates plotting, it would allow the visualization of St. Himark’s location layout.
Issue2.2.png


Issue3.png


Issue Faced 3
Issue: During the process of filtering the shake intensity by day, the time field provided by timestamp have been wrongly identified. Tableau has taken then date 04/08/2020 and 04/09/2020 as 4th august and 4th September instead of 8 and 9 of April. If the issue is not identified, the analysis of the earthquake will be identified as separate events rather than one event resulting in a misrepresentation and inaccurate Visual Analysis
Solution: To resolve this issue, time is being split into Date and time using excel’s Data, text to column function to separate the timestamp and changing the Date format into type: d-mmm-yy.

Data set Import Structure & Process

With the data-set analysis and transformation phase completed, the following files will have to be imported into Tableau for analysis:

Import.png

Each of the data file is added as a data source in Tableau. The relationship defined between each data source is the id or locations in the city. This will allow analysis to be conducted across all the data sources.

Additional processing was perform on both data sources:

  • Import the Damage and Intensity Report Data
  • Add new data connections to the Damage and Intensity Report Data. The new data connection file will be St. Himark’s Geometry Shape file
  • Perform inner join between both data sources using Damage and Intensity Report Data’s Location and St. Himark shape file’s ID. This will create a relationship between the Geometry provided with the locations. The following shows the configuration of the join:
Import1.png

Interactive Visualization

The interactive visualization can be accessed here: https://public.tableau.com/profile/anne1684#!/vizhome/Assignment1_15703574756750/Home?publish=yes

For the best experience, adjust your screen resolution to full screen by enabling full screen on the browser. Throughout all the different dashboards, useful guides and tips are provided to help users navigate through the different filters and actions to increase the efficiency of their analysis. The following elements are used through out all the dashboards to maintain consistency:

Interactive Technique Rationale Brief Implementation Steps

Filter dates with the use of time range slider

IV1.png

To provide flexibility for analysts to choose the time period which they are interest in analyzing. The use of checkbox or radio boxes is time-consuming as it requires analysts to check or unchecked the box.

  1. There should be a date field with its data type set to “date”
  2. Add the new field to be filtered.

Filter data source by Area types using a single selection drop down list

IV2.png

To allow analysts to concentrate on the data collected from each Areas with the use of a single selection. The use of a drop down-list also allows analysts to easily view all the selection choices and select the area type that they are interested to analyse.

  1. Adding the Area Dimension into filter field
  2. Configure the filter selection to be single selection drop down list

Filter Locations with the use of range slider

IV3.png

To provide flexibility for analysts to select a few locations that they are interest to analyse.

  1. Adding the location Measures into the filter field
  2. Configure the filter selection to be multi-selection slider range

Home Dashboard

To have an in-depth analysis on all the data attributes provided, it will not be possible to display all the attributes in a single dashboard. Hence, 3 separate dashboards (excluding home page) were created to provide a detailed visualization

Dashboard1.1.png

To allow flexibility and easy to understand in navigation, the following interactive techniques and instructions have been employed:

Interactive Technique Rationale Brief Implementation Steps

Navigate across dashboards with buttons


To provide users with the flexibility of moving across dashboards using a simple and easy to understand user interface.

  1. Create image buttons by dragging the button object into the dashboard
  2. Edit the button style to Image button and choose the button image used.
  3. Select the page to navigate to

Display tool-tips when user hovers over each button


To provide users with contextual information about the dashboard and the expected information and chart they will look at after selection.

  1. Edit the image button in the dashboard
  2. Under the tool-tip, enter the Tool-tip message to display when user hovers over the button.

Dispatch Priority Dashboard

The purpose of the Dashboard is to allow analysts to identify Locations that has reported high Shake Intensity and Damage. Additional Facilities Annotations was added to display crucial information that are useful for analysis and visualization. The following shows the Damage and and Shake Intensity Dashboard of St. Himark.

Dashboard2.1.png

To allow flexibility and easy to understand in navigation, the following interactive techniques and instructions have been employed:

Interactive Technique Rationale Brief Implementation Steps

Filter dates with the use of time range slider

IV1.png

To provide flexibility for analysts to choose the time period which they are interest in analyzing. The use of checkbox or radio boxes is time-consuming as it requires analysts to check or unchecked the box.

  1. There should be a date field with its data type set to “date”
  2. Add the new field to be filtered.

Filter data source by Area types using a single selection drop down list

IV2.png

To allow analysts to concentrate on the data collected from each Areas with the use of a single selection. The use of a drop down-list also allows analysts to easily view all the selection choices and select the area type that they are interested to analyse.

  1. Adding the Area Dimension into filter field
  2. Configure the filter selection to be single selection drop down list

Highlight Locations upon hover


When analysts hovers over the specific location on the map, the tool-tip displays the differences in Shake Intensity and Damage reported. This provides visual aid to the analysts.

  1. Under Dashboard > Actions, add an action called highlight.
  2. Select the source sheet, target sheets and target highlighting fields

Additional information displayed on Geometric map

IV4.png

To provide users with additional visual and contextual information to aid with their analysis.

  1. Annotate Area and entering the information to be displayed.

Hover over Geographical Map to specific data points to display the differences in Shake Intensity and Damage reported

IV4.1.png

To provide users with additional visual and contextual information to aid with their analysis.

  1. Select Tool-top > Insert > Sheets and select the sheet to be displayed in the tool-tip when hovered.
IV4.2.png

Earthquake Occurrence Dashboard

The purpose of the Dashboard is to provide an overview of the Shake Intensity and Damage across all 6 days. Additional charts below are displayed to show an in-depth analysis of Earthquake's Shake Intensity and Damage by Day, Hours and Location. Filters are consistently placed at the right side to increase usability. The following shows the date, time and location in which Earthquake had occurred.


Dashboard 3.png

To allow flexibility and easy to understand in navigation, the following interactive techniques and instructions have been employed:

Interactive Technique Rationale Brief Implementation Steps

Filter data source by Area types using a single selection drop down list

IV5.png

To allow analysts to concentrate on the data collected from each Areas with the use of a single selection. The use of a drop downlist also allows analysts to easily view all the selection choices and select the area type that they are interested to analyse.

  1. Adding the Area Dimension into filter field
  2. Configure the filter selection to be single selection drop down list

Filter data source by Date using a single selection slider

IV6.png

To enable analysts to view the damage or Shake intensity by specific day and to identify hidden insights.

  1. Add the Date Dimension into filter field and selecting view by day

Image Buttons by Damage or Shake Intensity

IV7.png

To allow analysts to concentrate on the data collected a single information between Damage and Shake Intensity

  1. Create image buttons by dragging the button object into the dashboard
  2. Edit the button style to Image button and choose the button image used.
  3. Select the page to navigate to

Filter across all the charts based on the one chart


To allow analysts to view the information in detail and to see how one data point is correlated to another data point in another chart given the same time.

  1. Under Dashboard > Actions, add an action called filter.
  2. Select the source sheet as Date of Earthquake
  3. Target sheets as Damage and location of earth quake

Hover over data points of each boxs to display more information about the hour of the day and Shake Intensity or Damage

IV7.1.png

When analysts is interested in a specific data point, they can simply hover over it and Tool-tip displays relevant details which might be crucial for analysis

  1. Drag the relevant fields into the Tool-tip
  2. Format the Tool-tip according to preferences

Uncertainties Analysis Dashboard

The purpose of the Dashboard is to allow analysts to detect the sudden changes in the number of responses sent over the period of time. Using the charts to identify uncertainties in data reported and relationship and changes in reporting over time. The following shows a detailed records of Number of response received throughout the 6 days of Earthquake.

Dashboard4.png

To allow flexibility and easy to understand in navigation, the following interactive techniques and instructions have been employed:

Interactive Technique Rationale Brief Implementation Steps

Filter Locations with the use of multi-selection drop-down checkbox

IV8.png

To provide flexibilty for analysts to select a few locations that they are interest to analyse.

  1. Adding the location Measures into the filter field
  2. Configure the filter selection to be multi-selection drop down list

Filter data source by Date using a single selection slider

IV1.png

To enable analysts to view the damage or Shake intensity by specific day and to identify hidden insights.

  1. Add the Date Dimension into filter field and selecting view by day

Additional information display on dashboard to explain purpose of dashboard

IV9.png

To provide users with additional visual and contextual information to aid with their analysis.

  1. Drag Text object into the dashboard and enter relevant information related to dashboard

Filter data source by Area types using a multi-selection check-boxes

IV10.png

To allow analysts to concentrate on the data collected from each Areas with the use of a single selection. The use of a drop down-list also allows analysts to easily view all the selection choices and select the area type that they are interested to analyse.

  1. Adding the Area Dimension into filter field
  2. Configure the filter selection to be a multi-selection check boxes

Interesting & Anomalous Observations

Using the dashboard as a platform for investigation and analysis, the following aims to provide answers to the question posed.

Q1: Emergency Responders’ prioritzed neighbouroods response based on Visualized damaged reports and parts of that City that are hardest hit.
Patterns Visualization
1.1 Locations with high reported Damage and Shake Intensity

Based on the Geographical Image provided, it depicts the North-East side of the City to be located closes to the Earthquake Zone. But in contrast to the image, Locations in South East Side were shown to reflect a higher damage as compared to Locations in the North East side.

Figure 1. shows the damage reported by each of the locations.

Figure 1.1. shows the scatter plot of the level of priority according to the percentile of damage and shake intensity reported.

According to both Figures, it shows that locations 3, 7, 8, 9 and 10 has one of the highest damage reported.

Further analysis was done to identify the reason for difference in Location affected as shown in the Geographic image and the reported Locations.

Observation1.png

1.2 Area with high reported Damage and Shake Intensity

Note: Area type includes Building,Medical, Power, Roads and Bridges

Filtering the Map Overview by Area type, it can be identified that the reason for high damage reported in locations 7, 8, 9, 10 and 11 might be due to a high number of buildings, power plants and Road and Bridges which are prone to damage due to Earthquakes.

Based on the data, we can infer that Emergency Responders should prioritize areas with a higher number of Buildings, Power plants and Road and Bridges.

Based on the City description provided, it is essential that emergency responders monitor location 4 closely, as Nuclear Power Plants are vulnerable to earthquake and may potentially impact people over a large area causing damage by fires, explosion and release of radioactive materials.

In addition to the potential harm, Residents in location 3 consists of mostly older-single family who lives close to the Nuclear Power Plant. In case of Earthquake, the residents might not be able to evacuate quickly and might require addition assistance. If no efficient evacuation is provided, Location 3, SAFE TOWN, might face the highest casualty.

Observation1.1.png
Q2: Use visual analytics to show uncertainty in the data. Compare the reliability of neighborhood reports and identify which neighborhoods are providing reliable reports.
Patterns Visualization
2.1 Date and Time of Earthquake and Pre-Shake Occurrence

Figure 2.1 provides the information on Date and Time in which the Earthquake occurred.

Inferring from figure 2.1, dates 8, 9 and 10 has showed signs of Earthquake occurrence as it has a higher shake intensity being reported.

Note: Hour of Time ranges from 0 to 23

On April 8, Earthquake was detected to have happened at 10am in the day.

On April 9, Earthquake was detected to have happened at 2am in the night.

Lastly, on April 10, an Earthquake was detected to have happened at 11pm in the afternoon.

After doing a more in-depth drill down using Damage and Shake Intensity by day, a pre-shake was identified to have occurred on April 6 from 3pm to 7pm.

Observation2.1.png
2.2 Location of Earthquake Occurrence

According to the pre-shake image provided, it was shown that the Earthquake happened on the North-East side of the city. Using the Geographic map, the areas that likely to be affected by the Earthquake are 2, 3, 4, 12, 14 and 7.

In contrast to the image, based on the Location of Earthquake Occurrence by Damage and Shake Intensity as shown in Figure 2.2, locations 3, 8, 9 and 14 were seen to have a high shake Intensity and Damage.


Observation2.2.png
2.3 Uncertainties in data reported

It is seen that some data reported might not be accurate as there might be power outrages in the specific areas resulting in a lost of response or delayed in response which will result in sudden drop or spike in number of response.

Based on Figure 2.3, Location 3 is identified to have unreliable data. There are frequent power outrages which resulted in a delayed response time and a sudden spike in the number of records at the later time of the day. According to City description, location 3 is identified to be doing a modernization to their electrical distribution system which might have contributed to the outrages.

Based on Figure 2.3, Location 4 is identified to have frequent power outrages which results in the lost of data which might be why the Shake Intensity report was low despite located near the Earthquake Zone.

Based on Figures 2.3.1 and 2.3.2, Locations 8, 9, 10, 11 and 17 are identified to have unreliable data as there are lost of response data or delayed response time due to the Earthquake, resulting in the spike in number of response at a later time.

Overall, Locations 3,4,8,9,10,11 and 17 were found to be unreliable.

Observation2.3.png
Observation2.3.1.png
Q3: Identify how conditions change and uncertainty change over time.
Patterns Visualization
3.1 Changes in Conditions

Based on Figure 3.1, It can be inferred that the number of response is closely related to the shake intensity.

Seen in day 6 and 7, despite having a high damage, when the shake intensity are low, the number of response are low as well. As the shake intensity increases, as seen in day 8 and 9, the number of response is seen to have a sharp increase.

On days 10 and 11, the response had a huge drop and came to a sudden stop when the Earthquake has ended.

Viewing by damage as shown in Figure 3.1.1, it can be seen that there was a sudden increase in reports of high damage at the second day. Using the happenings of Earthquake shown in Figure 2, the report is identified to be inaccurate as Earthquake occurrence had identified to be on the 3 day.

From both information, it can be inferred that on day 2, after Earthquake was announce, citizen started to panic and gave a report higher than actual, resulting in inaccurate reports..

Subsequent days, 3,4 and 5, it started to provide a more accurate representation.

Observation3.1.png
Observation3.2.png
3.2 Uncertainties changes over time

Based on figure 3.3, it can be inferred that during the occurrence of Earthquake, there was an increase in number of power outrages which resulted in lost of data and delayed response, making the information collected less reliable.

On days 6 and 7, although there are occasional lost of data, the number of response are consistent. Whereas days 8 and 9, on specific hours there was sudden increase in response and sudden lost of response. From which, we could infer that on days 8 and 9 the reports are less reliable. As response are sent and received at the different timings from the actual situation. While day 10 it starts to stabilize and eventually people stopped providing response as the Earthquake has stopped.

Observation3.3.png

References

https://wiki.smu.edu.sg/1617t3isss608g1/ISSS608_2016-17_T3_Assign_GUAN_YIFEI_Visualization

https://wiki.smu.edu.sg/1617t1IS428g1/IS428_2016-17_Term1_Assign3_Gwendoline_Tan_Wan_Xin

https://wiki.smu.edu.sg/1617t1IS428g1/IS428_2016-17_Term1_Assign3_Lim_Kim_Yong

https://wiki.smu.edu.sg/1617t1IS428g1/IS428_2016-17_Term1_Assign3_Tan_Kee_Hock

https://vast-challenge.github.io/2019/MC1.html

https://public.tableau.com/profile/gwennisme#!/vizhome/Assignment3_145/Home

https://en.wikipedia.org/wiki/Help:Table#Color;_scope_of_parameters