TrajAnalytics User Guide

User Manual:

Open the PDF directly: View PDF .
Page Count: 26

Download
Open PDF In Browser	View PDF

Table of Contents
Introduction ......................................................................................... 错误!未定义书签。

I.
1.1

Aims of NeighborVis ....................................................................................................... 2

1.2

NeighborVis Features ..................................................................................................... 2

II.

GSE Datasets ........................................................................................................................ 3

III.

NeighborVis Overview ......................................................................................................... 4

3.1 Overview of NeighborVis Data Operations.......................................................................... 4
3.2 Overview of NeighborVis Interface ...................................................................................... 4
IV.

NeighborVis Views and Functions ....................................................................................... 5

V.

Usage Example ...................................................................................................................19

VI.

Summary ............................................................................................................................25

•

NeighborVis Website: http://vis.cs.kent.edu

•

Software Development Team and Contact Information
Director:
Ye Zhao, Professor
Phone: 330-672-9059 Email: zhao@cs.kent.edu
Dept. of Computer Science, Kent State University, Ohio 44242
Co-Directors:
Andrew Curtis, Professor, Dept. of Geography, Kent State University
Xinyue Ye, Associate Professor, Dept. of Geography, Kent State University
Jing Yang, Professor, Dept. of Computer Science, UNC Charlotte
Lead of Software Development:
Chao Ma, Dept. of Computer Science, Kent State University
Other Team members:
Shamal AL-Dohuki, Farah Kawm, Suphanut Jamonnak

•

This work is partially supported by U.S. National Science Foundation under grant
number 1637242

1
NeighborVis Software User Guide

I. Introduction
1.1 Aims of NeighborVis
NeighborVis is a visual analytics system of geospatial-semantic event data in localized
neighborhoods within a larger city, town, suburb or rural area. The geospatial-semantic event
data (we name it GSE data in this document), consists of geospatial (longitude/latitude or
locational address), time, and semantic dimensions, which forms a large part of the important
information people deal with for business, personal and administrative use. Here are some GSE
data examples:
•

Social media data of geo-tagged twitter, facebook, Instagram posts;

•

Real estate data with house price, buyer information, loan and other related documents;

•

Police data of crime reports with crime locations, involved personnel, and case details;

•

Public health data of epidemic disease cases with locations and descriptions;

NeighborVis is developed to allow users to study these GSE datasets in a community
neighborhood. It is developed to answer the simple questions about a community, such as: what
is happening at this location, is it good or bad, and can outcomes be improved? It may be a police
department considering a blighted street, a planning department considering urban renewal and
the “greening” of spaces, a community group wanting to add neighborhood gardens, or a hospital
mining its child injury data. A large variety of users, such as real estate analyzers, public health
and safety administrators, business planners, and social scientists, are enabled to conduct
exploratory study through an integrated interface of maps, charts, semantic, and information
visualization tools, while a set of fluent user interactions are incorporated around the way user
think, feel and behave. The visual analysis is easily conducted over fine-grained geographical
resolutions such as small neighborhoods around local restaurants, churches, or schools with the
provisioning of temporal, semantic and geographical filters and constraints. NeighborVis can
greatly help users from multiple sectors including business, finance, public services, social
research, and local governments by leveraging the advanced computation and visualization
capabilities to utilize their large, complex, and rapidly accumulating community data.

1.2 NeighborVis Features
NeighborVis is developed as a Web-based system so that users can use it through web access
to remote server platforms. It can also be used solely as a localized version on desktops and
laptops. As illustrated by Figure 1, it supports a full data processing pipeline with the functions of:

2
NeighborVis Software User Guide

•

Data Loading and Processing: allow users to upload raw data including GSE data and
geographical shape files. The loaded data NeighborVis supports users to upload multiple
datasets.

•

Database Storage: The loaded GSE datasets are transformed and stored into the
NeighborVis database. The corresponding map with geocontext is also loaded. The
database we currently used is designed based on MongoDB and it supports both local and
server based operations.

•

Data Management: allow users to interactively define active GSE data resources for visual
analysis. Users can select more than one datasets at the same time which is important for
comparable and integrative study of multiple data sources.

•

Visual analysis: allow users to conduct interactive data visualization and analysis with a
set of charts, filters, maps, and other visual functions through human-machine
interactions.
Data Loading

Database Storage

GSE Data 1

Data Upload

Interactive Data
Query and
Management
Region Query

GSE Data 2

NeighborVis
Interface

User Shapefile
Query

...

Data
Preprocessing

GSE Data k

NeighborVis

Keyword Query

DataBase

Time Constraints

Load Map

Map
Visualization

Multi-Data
Management

Information
Visualization

Figure 1: NeighborVis system.

II. GSE Datasets
GSE data consist of a set of geospatial semantic events. Each event at least contains three
important attributes, location, time, and semantic description, while many of them have a
categorical attribute (about classes, types, or kinds) as well. Then the data may have more fields
related to specific fields but cannot be predefined in NeighborVis. In data loading function, we
check and validate that each raw dataset has the following attributes:
(1) Event ID, a unique identification number for a GSE record.
(2) Date/Time, which indicates when this event happened. This attribute should have the
format as YYYY-MM-DD HH:MM:SS.
(3) Geolocation of longitude and latitude, which indicates where an event happened.
(4) Text, which is the semantic description of this event.
(5) Category (optional), which is the category this event belongs to.
3
NeighborVis Software User Guide

III. NeighborVis Overview
3.1 Overview of NeighborVis Data Operations
GSE Data Preparation

GSE Data Queries

Data Loading

Working Data
Selection

NeighborVis
database

Working
Datasets

Check/
Uncheck

Active
Datasets
Global Event View
Heatmap, Markers

GSE Data Analysis

Query Selection
by Regions/Shapefiles

Map View of
Queries

Query
Management

Visual Analysis
of Queries

Figure 2: NeighborVis Data and Operation Flow.

Figure 2 shows the data and operation flow of NeighborVis system. The use of NeighborVis
consists of three stages:
1. GSE Data Preparation
Raw data is first loaded into NeighborVis database. Then users can select a set of working
datasets for their current task.
Once the data preparation is done, the working datasets can be interactively checked to
make as active datasets. Users can view events of active datasets on the map globally as
heatmap or markers.
2. GSE Data Queries
A query is the fundamental operating units of NeighborVis analysis. One query refers to a
set of events (may coming from multiple active datasets) inside a geospatial region and a
given time period. Users can define a query by drawing regions on the map by mouse, or
loading predefined regions from GIS shapefiles.
3. GSE Data Analysis
Once a group of queries are defined, users can manage them to visualize one or more
queries on the map. Meanwhile, details of queries are shown to study their contents. A set
of visual representations and tools are implemented to support interactive analysis of
queries.

3.2 Overview of NeighborVis Interface
Figure 3 shows the overview of NeighborVis interface. It includes a set of views to implement
the functions of GSE data loading, query, and visual analysis:
A.
B.
C.
D.
E.

Data Loading
Datasets Management Panel
Query Selection Panel
Query Management Panel
Map view
4
NeighborVis Software User Guide

F. Query Analysis view

Figure 3: NeighborVis System Interface.

Next, we describe these functions with a full user guide of operations.

IV. NeighborVis Views and Functions
A. Data Loading
Users load raw data to NeighborVis database. NeighborVis supports users to upload multiple
datasets in CSV format. A dataset is checked to make sure it includes the necessary data
attributes (as explained in Chapter II). Users are required to specify the correct attributes in the
CSV file. Then, the system automatically generates geoindex and stores the dataset in database.
NeighborVis database is specifically designed and implemented over the MongoDB database.
Click Data Manager will start new data uploading process.
1. Start data uploading
Click Data Manager, the popup window shows all the available datasets in NeighborVis
database. Click Data Upload button to start uploading new data.

5
NeighborVis Software User Guide

2. Name your data
You are prompted to give a name of the new dataset. The name should be unique, which is
different from existing datasets.

3. Select your data file in CSV format. The file will be uploaded.

6
NeighborVis Software User Guide

4. Specify required data attributes (column names) in the uploaded CSV file, including ID,
longitude, latitude, time/data, text description, and categories(classes).

7
NeighborVis Software User Guide

5. Convert the CSV data into NeighborVis GSE tables. The system automatically completes
geoindexing

6. Finish data uploading process

7. After uploading, the newly added dataset automatically appear on the Datasets
Management Panel.
8
NeighborVis Software User Guide

B. Datasets Management Panel
1. Show and hide datasets from the panel
Datasets management panel shows a set of data objects, which are NeighborVis datasets
users are currently working on. By clicking Data Manager button, in the popup window of
Figure 4, users can select which datasets (already uploaded to NeighborVis) are shown in
this panel, which are working datasets. Other datasets are hidden from this panel, if it is not
being used in current analysis tasks.

Figure 4: Data manager interface
2. Delete datasets
Users can click delete icon, which fully delete the corresponding dataset from NeighborVis
database and system. You will need to upload a dataset again once fully deleted.

Figure 5: Datasets Management Panel.

3. Check/uncheck active datasets
On the panel shown in Figure 5, user can check or uncheck a working dataset to make it an
active dataset. The query selection (see next section) is effective on all active datasets.
4. Show events of active datasets
Users can operate on the panel to globally show all events on the map view of specified
datasets.
9
NeighborVis Software User Guide

Click to show/hide all events from the corresponding dataset as markers on the map
Click to select color of the markers shown on the map
Click to show/hide all events from the corresponding dataset as heatmap on the map

C. Query Selection Panel
Figure 6 shows the Query Selection Panel interface. It allows users to set up multiple
conditions of query generation.
First, the selected data list shows the names of active datasets, which user checked on the
Datasets Management Panel.
Then, users set up filters of
(1) Time period
(2) Keywords in text description field: The keywords can be combined by Boolean
operators: AND OR, match such as “sandy” and “floor”, “sandy” or “floor”.
a. Fuzzy matching is also supported such as “sand” -> “sandy”; “floo” ->“floor”.
(3) Category name: The input must be one of the names in the origin datasets.

Figure 6: Query Selection Panel

After these query filters are set, users can define a region of interest by drawing on the
map. Click the icons on the map will allow users to select regions (See IV.E for details),
the system will then generate a query. The query will include GSE data based on the
region and the filtering conditions specified here.

10
NeighborVis Software User Guide

D. Query Management Panel
A query includes data of
(1)
(2)
(3)
(4)

Spatial region;
Time period;
Keywords and category based sementic filters;
GSE data inside the region, within the time period, and satisfy the semantic filters.

Figure 7: Query Management Panel

All the generated queries are managed in the Query Management Panel, as illustrated in
Figure 7. The list shows the queries, how many active datasets are involved, and how many
data records (events) are included.
Users can delete a query on the panel by clicking the garbage icon.
Users can click any query to make it active, so that the map view and visual analysis view
show the corresponding query data for analysis.
A query may be applied over multiple datasets. Users can also select one specific dataset
inside the active query to highlight it on the map view. In default, all datasets are assumed
to be used in visualization.

E. Map view
Map view allows users to select regions for queries, display query regions, display GSE data
on the map. Figure 8 shows one snapshot of the map view, where one active query (the
rectangular region) is selected. The internal markers show the GSE events data of this query.
They come from three different datasets (i.e., the checked data objects in the datasets
11
NeighborVis Software User Guide

management panel). Users can click to select one marker to display its details. The legend
on the lower left corner displays the colors of markers representing these datasets, and the
number shows the number of markers (events).
On the map view ,multiple queries can be displayed simultaneously, while one active query
is highlighted which is user selected on the query management panel.

GPS Location of
The Selected
Marker (Event)

Popup Window of
Data manager

Map Icons
Detail View of the
Selected Marker
(Event)

Map

One Query Region
With Markers of
GSE Events

Legend: Markers
Color and the
Number of Events of
Different Datasets

Figure 8: Map view overview

1. Map Icons and Functions

Zoom in and out
Select map types

Select polygon region
Select rectangular region
Select circular region
Relocate map position
Load shapefile

Figure 9: Map icons and their functions

12
NeighborVis Software User Guide

Figure 8 shows the icons on the map view which also indicates the functions each icon
represents including
(1) Query region selection types: allow users to select regions of interest to define a new query;
(2) Load pre-defined shapefiles to to define one or multiple queries;
(3) Map operations of zooming, map styles, and relocate the map position (if necessary when
you change a different GSE dataset).
2. Map View Types
(1) Show each GSE data event as point markers

Users can click button
the events on the map.

for each dataset on Datasets Management Panel to globally show

(2) Show GSE data distribution as heatmap

13
NeighborVis Software User Guide

Users can click button
for each dataset on Datasets Management Panel to globally
show the event distribution on the map.
3. Show GSE Event Details
(1) Users select one marker to show the event details in a popup window.

14
NeighborVis Software User Guide

(2) Many GSE events happen at the same location. When users click the location, a time line
view is popped up to show these events happening along a time axis. Users can
interactively zoom in on the time line to see the time varied events. Users can click one
event to see its details as well.

F. Query Analysis View
Query Analysis View is a panel on the right side of the map, which is initially hidden. Users
can open it to visualize the details of the current active query. If users select one dataset of
this query, only events related to this selected dataset is applied (see IV.D for details).
Otherwise, all events in this query is visualized.
There are four sections in the Query Analysis View including
(1) Event records list.
(2) Time bar chart.
(3) Keyword cloud.
(4) Category bar chart.
All these components are interactively operated by users to filter by time, keywords and
categories. The visualizations over these components, as well as on the map, are
coordinated and synchronized with users’ operations.

15
NeighborVis Software User Guide

1. Event Records List
It shows all events in the active query and the active dataset. Users can hover over each
item to show it on the map view.

2. Time Bar Chart
It shows the time distribution of all events in the given time period. Users can click to
choose one particular bar so that only events inside this time range are used in map
view and in the other views of this analysis panel. Users can choose to show the time
bar charts in yearly, monthly or daily manner, according to the nature of the GSE events.

16
NeighborVis Software User Guide

3. Keyword Cloud
It displays the keywords in the descriptions of all events of the query. The font of each
keyword is used to visualize the importance of this keyword, currently measured by the
term frequency. Users can click on a keyword so that the events on the map and on this
panel will be filtered based on this selection.

.

17
NeighborVis Software User Guide

4. Category Bar Chart
It shows the histogram of events on the event categories. Users can click on a bar so
that the events on the map and on this panel will be filtered based on this selection.

Figure 8: (3) Category bar chart shows top categories.

5. Interaction
All these charts and keyword cloud are interactively operated by users to filter this query by
different conditions. These conditions can be combined. For each component, the condition
filter is shown on the top bar. Users can click on the check to remove this condition.

18
NeighborVis Software User Guide

Meanwhile, all these conditions are shown in the top of the Query Analysis View as an
information box. Map view and this Query Analysis View are updated automatically in real
time based on the changes of these conditions.

V. Usage Example
We show one example of using NeighborVis to analyze a real-world data. This is a geo-tagged
social media data which is collected and processed from Twitter API. The data is extracted for
the area in Brooklyn, NY, during the time period of Hurricane Sandy 2012, which created
particularly severe damage in New Jersey and New York of USA. All the events are divided into
three GSE datasets. Each is a part of all the tweet events related to Sandy and has geo-tags
(locations). Three of them refer to the time periods before, during, and post Hurricane Sandy,
respectively. Next we show how NeighborVis is used in visual analysis.

1. Upload a dataset and name it “sandypost”. Other two datasets are uploaded in the
same way (SandyBefore and SandyDuring).

19
NeighborVis Software User Guide

2. All three datasets are stored in the database and selected as working datasets
(SandyBefore, SandyPost and SandyDuring). They are all shown in the Datasets
Management Panel.

3. Relocate the map to the corresponding area (Brooklyn). Set color for each dataset.

4. Check to see heatmaps (or markers) of active datasets on the map. Two datasets
(SandyPost and SandyDuring) are activated for the following query.
20
NeighborVis Software User Guide

5. Set up query selection filters by giving a time period from 2012-10-25 to 2012-11-03,
the keyword of interest “sandy”.

6. Select map icons to draw a region on the map. All events satisfy the filters and inside
this region are used to form one active query. That is: all GSE events (geo-tagged
tweets) from 2012-10-25 to 2012-11-03 and contain keyword “sandy”. This query also
21
NeighborVis Software User Guide

appears in the Query Manager Panel. Other events out of this query are shown in
small dots as background. Users can make them disappear by operating on the above
panel as well.

7. Perform two more queries. These queries are shown together on the map with Query
1 being highlighted. Users can select different active queries on the Query Manager.

22
NeighborVis Software User Guide

8. By clicking “Query 1” in the Query Manager, a sublist shows the two datasets
(SandyPost SandyDuring) associated with this query. By selecting one of them
(SandyDuring), the Map View and Query Analysis View are updated to show only the
events in this specified dataset (SandyDuring).

9. Study Query Analysis View for the list of events, the time distribution, keyword clouds
and category distributions. Users can check event details on the map as well.

23
NeighborVis Software User Guide

10. Show time line details of events at a specific location, which may be the same person
published multiple tweets.

24
NeighborVis Software User Guide

VI. Summary
NeighborVis software is used for analyzing geospatial temporal events with semantic contents.
NeighborVis has two ways to access:
➢ Downloadable version to be used in desktop/laptop, please download it from our website
and refer to the installation guide for technical details.
➢ Internet-based Visual analysis through Web browsers, please go to the website for details

NeighborVis will be updated with more functions in both data processing and visualization.
Please help to provide your feedback, suggestions and comments on software website to help
us improve the software.

25
NeighborVis Software User Guide



---

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.7
Linearized                      : No
Page Count                      : 26
Language                        : en-US
Tagged PDF                      : Yes
XMP Toolkit                     : 3.1-701
Producer                        : Microsoft® Word 2016
Title                           : TrajAnalytics User Guide
Creator                         : TrajAnalytics Development Team
Description                     : A cloud visual analytics software
Creator Tool                    : Microsoft® Word 2016
Create Date                     : 2018:07:19 16:19:42-04:00
Modify Date                     : 2018:07:19 16:19:42-04:00
Document ID                     : uuid:333B63D2-05CE-4C12-B2E8-8D02ABE93BA9
Instance ID                     : uuid:333B63D2-05CE-4C12-B2E8-8D02ABE93BA9
Author                          : TrajAnalytics Development Team
Subject                         : A cloud visual analytics software

EXIF Metadata provided by EXIF.tools