/Hangzhou-Metro-Data-Visualizer

🚊 GUI implemented with Qt Creator

Primary LanguageC++

🚉 Hangzhou Metro Data Visualization

Environment: Windows 10--- Inplemented with Qt Creator

Introduction

image-20200402111627661

"Metro Insider"* handles files with SQLite in-memory database with high efficiency, the following modules are implemented:

-> Loading Files with a Selector and Filter

-> SQLite Runner

-> Plotting & Analyzing

-> Route Planning

Functions & Implementations Details

Module 1 -> Loading Files with a Selector and Filters

Select Files and Set Filters

"choose a folder" button is expected to be pushed to start the whole process. It will open a dialogue to let users to select the "dataset" folder. It is implemented with QFileDialog::getExistingDirectory().

image-20200402110928691

Then a tree widget is created to let users choose files and filters implemented with QTreeWidegetsItem.

-> Files Tree

The files tree can synchronize the child node with the parent node all the time, with Qt::Checked; Qt::PartlyChecked; Qt:Unchecked; which means that

  • parent Partlychecked $\Leftrightarrow$ some children Checked
  • parent Checked $\Leftrightarrow$ all children Checked
  • parent Unchecked $\Leftrightarrow$ all children Unchecked
  • It's implemented recursively.

And names of all the chosen files is maintained in QStringList chosen_files_names timely before

the "load chosen files" bottom being pushed. (So the program won't have any problems when users change the checked items during the loading process.)

-> Filters Tree

It is designed for choosing the fields to be loaded into the SQLite database. But "time" & "stationID" & "status" can not be unchecked since our mandatory task requires a plot on "Inflow & Outflow of a station".

Loading Selected Files

Only selected fields of selected files will be loaded into the datatbase. The database would create a table whose columns are those fields. And this module has the following characteristics:

-> High efficiency

Loading the dataset of one day takes averagely 50s in my surface-Laptop (2017, 13.5'), it's relatively fast due to the insertion of data to SQLite can't be done by multicore.

-> Multithreading

With <QtConcurrent> , the loading process is being done in another thread so the ui-thread will never be frozen.

-> Status Bar

The status bar is well designed to respond quickly to any situation

image-20191228185428567

image-20191228185746754

image-20191228185843719

If users change filters or files selections after a loading, the status bar will suggest a renew.

image-20191228190716316

-> Robustness
  • Loading process is based on a file list maintained before "pressing the load button", so even if users change the file selection box or filters during the load, they won't have a problem.
  • Reloading is well supported since the database will drop and create a new table newly designed for the chosen column and ensure there will not be a leak of memory.
  • Before the loading process have been done, all the buttons and combo boxes relied to chosen files and filters will be disabled to avoid possible errors.

Module 2 -> SQLite Runner

img SQLite Runner is designed for any explore of raw data by users. It is implemented by QTextEdit QSqlQueryModel QTableView , it has following characteristics:

-> Fast Speed
  • The database is in-memory and has a carefully designed composite index. So almost all the SELECT operation can be done in 0.x seconds.
  • QTableView with QSqlQuerymodelmakes it possible for dynamic loading, which means that the ui will not show all the results from e.g. SELECT * FROM TABLE at once so the ui will not freeze at any time. When users scrolls the table view, new query results will constantly be loaded
-> Syntax Error & Running Error Information

This function is implemented by

ui->statusbar->showMessage(model->lastError().databaseText(), 100000);

It will show any error message in the status bar.

Module3 -> Plotting & Analyzing

1585796701251

1585796708357

1585796715361