Difference between revisions of "IS428 2017-18 T1 Assign Dong Ruiyan"
(q1) |
(q1) |
||
| Line 5: | Line 5: | ||
=== Q1 === | === Q1 === | ||
''Characterise the sensors’ performance and operation. Are they all working properly at all times? Can you detect any unexpected behaviours of the sensors through analysing the readings they capture? Limit your response to no more than 9 images and 1000 words.'' | ''Characterise the sensors’ performance and operation. Are they all working properly at all times? Can you detect any unexpected behaviours of the sensors through analysing the readings they capture? Limit your response to no more than 9 images and 1000 words.'' | ||
| − | ==== | + | ==== Improperly Working Behaviours ==== |
===== Stop Working ===== | ===== Stop Working ===== | ||
The calendar chart shows how total number of records of all 9 monitors varies with the hour of the day over the 3 months. There are 5 blank girds, which shows the total number of records of all 9 monitors at the timestamps is 0. It strongly indicates that all 9 monitors stopped working at the timestamps period, which are 00:00 AM 2nd April, 00:00 AM 6th April, 00:00 AM 4th August, 00:00 AM 7th August and 00:00 AM 2nd December. There are 2 light green grids, which shows that the total number of records of all 9 monitors at the timestamps is much fewer than that at other timestamps. By viewing the calendar chart for different monitor (filter by monitor), only Monitor 3 was working and captured 2 records at 00:00 AM 2nd August, and only Monitor 7 and Monitor 8 was working and captured total 5 records at 00:00 AM 7th December. | The calendar chart shows how total number of records of all 9 monitors varies with the hour of the day over the 3 months. There are 5 blank girds, which shows the total number of records of all 9 monitors at the timestamps is 0. It strongly indicates that all 9 monitors stopped working at the timestamps period, which are 00:00 AM 2nd April, 00:00 AM 6th April, 00:00 AM 4th August, 00:00 AM 7th August and 00:00 AM 2nd December. There are 2 light green grids, which shows that the total number of records of all 9 monitors at the timestamps is much fewer than that at other timestamps. By viewing the calendar chart for different monitor (filter by monitor), only Monitor 3 was working and captured 2 records at 00:00 AM 2nd August, and only Monitor 7 and Monitor 8 was working and captured total 5 records at 00:00 AM 7th December. | ||
| Line 14: | Line 14: | ||
===== Exceptions ===== | ===== Exceptions ===== | ||
By viewing the line graphs of the hourly total readings of all chemicals per day by month captured by each monitor over the 3 months, I find that all monitors captured abnormally high readings. One clear example is the line graph of Monitor 9, which shows an extremely high reading of 46.91 at 03:00 AM 11th April that is much larger than the average readings of 2.53 at 03:00 AM across April. | By viewing the line graphs of the hourly total readings of all chemicals per day by month captured by each monitor over the 3 months, I find that all monitors captured abnormally high readings. One clear example is the line graph of Monitor 9, which shows an extremely high reading of 46.91 at 03:00 AM 11th April that is much larger than the average readings of 2.53 at 03:00 AM across April. | ||
| + | ==== Characterise Monitor Performance ==== | ||
| + | I characterise the monitors’ performance and operations based on the amount of noise in their readings. | ||
| + | ===== Large Amount of Noise ===== | ||
| + | ===== Medium Amount of Noise ===== | ||
| + | ===== Small Amount of Noise ===== | ||
| + | ===== Special: Monitor 4 ===== | ||
| + | The line graph shows the total readings of Monitor 4 at each timestamp over the 3 months. It shows a clear linearly-increasing in its readings over the time. However, other monitors did not show such kind of trend, and the possibility that this trend is caused by environmental change is very small. This trend might be caused due to errors or improper monitor behaviours. | ||
=== Q2 === | === Q2 === | ||
=== Q3 === | === Q3 === | ||
Revision as of 19:51, 8 October 2017
Visual Analytics Assignment
Contents
Problem & Motivation
Questions
Q1
Characterise the sensors’ performance and operation. Are they all working properly at all times? Can you detect any unexpected behaviours of the sensors through analysing the readings they capture? Limit your response to no more than 9 images and 1000 words.
Improperly Working Behaviours
Stop Working
The calendar chart shows how total number of records of all 9 monitors varies with the hour of the day over the 3 months. There are 5 blank girds, which shows the total number of records of all 9 monitors at the timestamps is 0. It strongly indicates that all 9 monitors stopped working at the timestamps period, which are 00:00 AM 2nd April, 00:00 AM 6th April, 00:00 AM 4th August, 00:00 AM 7th August and 00:00 AM 2nd December. There are 2 light green grids, which shows that the total number of records of all 9 monitors at the timestamps is much fewer than that at other timestamps. By viewing the calendar chart for different monitor (filter by monitor), only Monitor 3 was working and captured 2 records at 00:00 AM 2nd August, and only Monitor 7 and Monitor 8 was working and captured total 5 records at 00:00 AM 7th December.
Redundant Records
By viewing the chart of total number of records of each chemical at each timestamp for different monitor (filter by monitor), most of the monitors captured the amount of AGOC-3A released twice at multiple timestamps over the 3 months. The above chart is one example for Monitor 5. There are multiple dark green bars, which shows the total number of records of AGOC-3A at that timestamp is 2. It indicates that Monitor 5 captured the amount of AGOC-3A released at that timestamp twice. By viewing the chart for different monitor (filter by monitor), most of the monitors captured the amount of AGOC-3A released twice at multiple timestamps over the 3 months. Besides, in most cases of the duplicated records of AGOC-3A at a timestamp, the amount of AGOC-3A of one record is much higher than (around 10 times) that of the other record. (To view that, go to ..dashboard) Since the smaller readings are at the normal range while the larger readings are abnormally high, I assume that the large readings are outlier due to improper working behaviours of the monitors. The extremely high readings may dominate the following analysis. To avoid that, the records with smaller reading of AOGC-3A are kept and the records with larger reading will be ignored. To do that, I created a …
Exceptions
By viewing the line graphs of the hourly total readings of all chemicals per day by month captured by each monitor over the 3 months, I find that all monitors captured abnormally high readings. One clear example is the line graph of Monitor 9, which shows an extremely high reading of 46.91 at 03:00 AM 11th April that is much larger than the average readings of 2.53 at 03:00 AM across April.
Characterise Monitor Performance
I characterise the monitors’ performance and operations based on the amount of noise in their readings.
Large Amount of Noise
Medium Amount of Noise
Small Amount of Noise
Special: Monitor 4
The line graph shows the total readings of Monitor 4 at each timestamp over the 3 months. It shows a clear linearly-increasing in its readings over the time. However, other monitors did not show such kind of trend, and the possibility that this trend is caused by environmental change is very small. This trend might be caused due to errors or improper monitor behaviours.
