Scripts for calculating statistical data for word-aoi-groups.
These scripts were created to facilitate analysis of multi-word AOI groups in linguistic eyetracking data.
SMI beGaze automatically generates AOIs around each word of a text stimulus. Our researchers needed a tool to perform multiple pass analysis, regression analysis etc. for variable groupings of these AOIs. Later a variant of the script that works with exported data from Tobii Studio was added.
The two scripts are meant to be used as command-line tools. You can simply download the executables in the dist directory, which were generated using pyinstaller on a windows system running Python 3.7, or, if you want to run the scripts on other systems or if you need to modify the scripts for your own purposes, download the sources. They should run on a standard Python 3.7 installation.
smi_aoi_group.exe SOURCE
Source file is expected to be a SMI BeGaze export file (Metrics export - Event statistics - single), where each line represents a fixation event including word AOI information.
First execution of the tool creates a SOURCE.list file containing all AOIs in the data file, grouped by stimulus, like this:
Critical_T1_C1a, 1, José
Critical_T1_C1a, 2, y
Critical_T1_C1a, 3, Andrés
Critical_T1_C1a, 4, comen
Critical_T1_C1a, 5, muchos
Critical_T1_C1a, 6, dulces.
Critical_T1_C1a, 7, A
Critical_T1_C1a, 8, pesar
Critical_T1_C1a, 9, de
Critical_T1_C1a, 10, ello,
Critical_T1_C1a, 11, están
Critical_T1_C1a, 12, delgados.
To calculate group metrics, create a file named SOURCE.groups. Define your groups using the AOI indices listed in the SOURCE.list file, e.g. to analyze the sentence fragment "A pesar de ello," from the example above, the SOURCE.groups file should look like this:
Critical_T1_C1a 7,8,9,10
You can enter multiple lines in one SOURCE.groups file, which will be processed consecutively.
On second execution (if a valid SOURCE.groups file is present), the tool generates a SOURCE.stats file like this:
Stimulus AOI group Participant Total dwell time Mean dwelltime per word Total num fixation First fixation start First fixation duration First pass duration First pass mean dwelltime per word First pass num fixations Second pass start Second pass durations Second pass num fixations Re-readings duration Total Skip Num regressions into Sources regressions into Num regressions out of Targets regressions out of Trial no. Trial start time Group word count Group char count
Critical_T1_C1a 7,8,9,10 EXP_01_P_01 356.2 89.1 3 1216.0 134.0 356.2 89.1 3 0.0 0.0 0 0.0 0.0 FALSE 0.0 0 2 27036112.4 4 12
Critical_T1_C1a 7,8,9,10 EXP_01_P_02 1129.1 282.3 6 6763.7 352.1 880.9 220.2 4 12658.1 118.1 1 248.2 62.1 FALSE 2.0 18,15 0 4 1721543.4 4 12
Critical_T1_C1a 7,8,9,10 EXP_01_P_03 386.2 96.6 4 1274.5 70.0 386.2 96.6 4 0.0 0.0 0 0.0 0.0 FALSE 1.0 8 0 11 2425390.3 4 12
Critical_T1_C1a 7,8,9,10 EXP_01_P_04 126.1 31.5 1 1403.4 126.1 126.1 31.5 1 0.0 0.0 0 0.0 0.0 FALSE 0.0 0 15 902713.3 4 12
Critical_T1_C1a 7,8,9,10 EXP_01_P_06 700.1 175.0 4 1808.1 320.1 320.1 80.0 1 4055.2 222.0 2 380.0 95.0 FALSE 2.0 18,18 0 4 2565602.6 4 12
Critical_T1_C1a 7,8,9,10 EXP_01_P_07 448.2 112.0 4 1873.7 190.1 448.2 112.0 4 0.0 0.0 0 0.0 0.0 FALSE 0.0 0 2 4545946.0 4 12
Critical_T1_C1a 7,8,9,10 EXP_01_P_08 250.0 62.5 2 2443.8 104.0 250.0 62.5 2 0.0 0.0 0 0.0 0.0 FALSE 0.0 0 4 1799017.6 4 12
Critical_T1_C1a 7,8,9,10 EXP_01_P_09 334.0 83.5 4 1421.1 66.0 280.0 70.0 3 3107.9 54.0 1 54.0 13.5 FALSE 2.0 9,14 0 19 3867765.5 4 12
...
Additionally, a SOURCE.debug file is created with detailed information on the AOI matching and calculations, which can be consulted in case the stats file contains unplausible results:
Stimulus Critical_T1_C1a Participant EXP_01_P_01 Trial no. 2 Trial start time 27036112.4 Group ['7', '8', '9', '10'] A pesar de ello Group word count 4 Group char count 12
Fix 1: José 1 1.5 164.1
Fix 2: José 1 193.5 122.0
Fix 3: Andrés 3 357.7 126.4
Fix 4: comen 4 515.7 182.1
Fix 5: muchos 5 743.8 160.1
Fix 6: dulces. 6 941.9 240.1
Fix 7: A 7 1216.0 134.0 <-
Fix 8: pesar 8 1376.0 132.1 <-
Fix 9: ello, 10 1612.1 90.1 <-
Fix 10: están 11 1740.2 162.0 FIRST PASS END
Fix 11: delgados. 12 1942.2 220.1
Fix 12: Les 13 2202.3 78.1
Fix 13: Les 13 2312.4 96.0
Fix 14: gusta 14 2512.5 122.3
Fix 15: gusta 14 2660.5 154.1
Fix 16: mucho 15 2864.6 124.0
Fix 17: ver 16 3028.7 306.1
Fix 18: ver 16 3346.8 66.0
Fix 19: ver 16 3430.9 198.1
Fix 20: ver 16 3665.7 285.3
Fix 21: televisión. 18 3989.0 174.1
Last AOI index: 18, Total dwell time: 356.2, Num fixations: 3
...
Stimulus Critical_T1_C1a Participant EXP_01_P_06 Trial no. 4 Trial start time 2565602.6 Group ['7', '8', '9', '10'] A pesar de ello Group word count 4 Group char count 12
Fix 1: José 1 1.9 101.4
Fix 2: Andrés 3 197.4 152.1
Fix 3: comen 4 425.5 102.0
Fix 4: muchos 5 701.6 194.1
Fix 5: comen 4 963.8 232.1
Fix 6: dulces. 6 1243.9 176.1
Fix 7: dulces. 6 1624.0 148.1
Fix 8: pesar 8 1808.1 320.1 <-
Fix 9: están 11 2180.4 182.0 FIRST PASS END
Fix 10: delgados. 12 2402.4 198.1
Fix 11: delgados. 12 2790.6 358.1
Fix 12: mucho 15 3368.8 106.0
Fix 13: gusta 14 3520.9 104.0
Fix 14: ver 16 3669.0 114.1
Fix 15: televisión. 18 3851.1 160.0
Fix 16: A 7 4055.2 128.0 <-
Fix 17: dulces. 6 4279.2 354.3
Fix 18: pesar 8 4677.4 94.0 <-
Fix 19: la 17 4955.5 196.1 SECOND PASS END
Fix 20: televisión. 18 5179.7 214.2
Fix 21: pesar 8 5421.8 158.0 <-
Last AOI index: 8, Total dwell time: 700.1, Num fixations: 4
...
Basically the same tool, modified to work with TSV export files from Tobii Pro Studio.
- Takara Baumbach (takb)
This project is licensed under the GNU General Public License v3.0 - see the LICENSE file for details
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent