Entry Name: "Purdue-Tang-MC2"

VAST Challenge 2016
Mini-Challenge 2

Team Members:

Hui Tang, Purdue University, tang227@purdue.edu PRIMARY

Zheng Zhou, Purdue University, zhou85@purdue.edu

Shuang Wei, Purdue University, wei93@purdue.edu

Mingran Li, Purdue University, li1940@purdue.edu

Siyan Liu, Purdue University, liu1690@purdue.edu

Hsin-man Wu, Purdue University, wu949@purdue.edu

Xinghe Hu, Purdue University, hu264@purdue.edu

Yuankun Song, Purdue University, song340@purdue.edu

Dr. Yingjie (Victor) Chen, Computer Graphics Technology, Purdue University, victorchen@purdue.edu (supervising faculty)
Dr. Zhenyu (Cheryl) Qian, Interaction Design, Purdue University, qianz@purdue.edu (supervising faculty)

Student Team: YES

Tools Used:

Excel, jQuery, Google Charts (adopting D3.js)

Approximately how many hours were spent working on this submission in total?

160 hours

May we post your submission in the Visual Analytics Benchmark Repository after VAST Challenge 2016 is complete? YES

Video

https://youtu.be/PhVmMB0FYrk

https://va.tech.purdue.edu/VAST2016/zhengz/metacurve.wmv

The MetaCurve System

https://va.tech.purdue.edu/VAST2016/zhengz/

Questions

MC2.1 – What are the typical patterns visible in the prox card data? What does a typical day look like for GAStech employees?

Limit your response to no more than 6 images and 500 words.

1. According to the prox card data, employees’ typical working hours are from around 7 AM to around 5:30 PM. GAStech employees go to the company in the morning at around 7 AM and stay at their own offices until noon. Then they go to first floor, stay there for about one hour (may be outside of the building for lunch), and come back to work at around 1 PM and leave company at around 5:30 PM. In the following image, each blue patch represents the aggregate counts for 14 days that an employee shows up in a particular zone at 5-minute interval. A dark blue patch indicates the employee stays at this location at this time for more than 4 days. Lighter blue patch means the employee appears less often at 2 to 3 times. A light blue patch with colorful dot indicates only one appearance during the 14 days. The colors of dots represent different dates. However, for the employees from different department, prox data patterns vary. For employees in security department, their time schedules are fixed. Their schedules repeat every day and barely go to other places, which indicates that the company’s security is heavily relied on electronic equipment. This also implies that the employees in security department monitor the situation mostly in their offices.

Employees in administration department and executive department spend most of the day in their own offices. But they still go to other zones of the building occasionally. The pattern can be seen in the following images as many light blue patches are spreading all over the building other than their own offices (the orange lines).

 For employees from other departments such as Engineering, Information Technology, Facilities, and HR, their schedules are more flexible than employees in security department. They may arrive at or leave their offices a few minutes early or late. But they stay in their offices for most of the working hours.

2. We also find 14 people from engineering, facilities, information technology departments work from afternoon to midnight. Their patterns are consistent with their department patterns.

3. Emile Arpa (earpa001) and Varro Awelon (vawelon001) work from midnight till morning (around 12AM – 7AM).

4 .  Raye Paredes (rparade001), Dylan Scozzese (dscozzese001), and Chi Staley (cstaley001) always leave early around 3:30PM.

MC2.2 – Describe up to ten of the most interesting patterns you observe in the building data. Describe what is notable about the pattern and explain what you can about the significance of the pattern.

Limit your response to no more than 10 images and 1000 words.

1. The power of public areas is always on. The Equipment Power and the Light Power curves show that Floor 1 Zone 3 (Main Entrance), Floor 1 Zone 8A (Floor 1 Corridor), Floor 1 Zone 8B (Floor 1 Corridor), Floor 2 Zone 12A (Floor 2 Corridor), Floor 2 Zone 12B (Floor 2 Corridor), Floor 2 Zone 12C (Floor 2 Corridor), Floor 3 Zone 11A (Floor 2 Corridor), Floor 3 Zone 11B (Floor 3 Corridor), and Floor 3 Zone 11C (Floor 3 Corridor) are always up and running. For example, The Floor 2 Zone 12A Equipment Power and Lights Power are plotted as follows, all the other aforementioned public zones are identical to this image except for the actual Watt readings.

2. The power of the server on floor 3 (The majority of HVAC floor 3 zone 9) is always up and running. We noticed this because for every other 35 HVAC zones, the Equipment Power and the Light Power in the same zone have downtime and have exactly identical on/off pattern (and therefore the aggregated outlier pattern pie chart on the right) during the 14 days.

3. There are regular weekday meetings/trainings holding in the Room 1030 (HVAC zone F_1_Z_5) and Room 2700 (HVAC zone F_2_Z_14) as the light and equipment power is on for a regular period during the weekdays. No meetings during weekend. The typical meeting time slots for Room 1030 are around 2:30 AM-3:30 AM, 6:00 AM-7:00 AM, 11:40 AM-1:40 PM. The typical meeting time slots for Room 2700 are around 10:30 AM-11:30 AM, 2:30 PM-3:30 PM, 6:30 PM-7:30 PM, and 10:00PM-11:00 PM.

4. The HVAC system is operating 24/7. The regular “1 hour on, 5 min off” periodic pattern appears in many building sensors. For zones in floor 1 and floor 2, this can be verified by looking at the reheat damper position, supply inlet mass flow rate, supply inlet temperature rises, and thermostat temperature. All of these readings above have drastic changes with same phase with the coil power. Moreover, we discover the same pattern for the VAV system of the entire floor.

5. Only for floor 1 and floor 2, the VAV system cooling coils and reheat coil of every zone observe this pattern of “1 hour On, 5 min Off”. Not for floor 3.

6. According to an almost consistently unchanged pattern in each HVAC zone, the thermostat temperature, the thermostat cooling and heating temperature setpoint in each zone are highly linked, although the specific setting among zones may be slightly different. This pattern shows as the cooling setpoint is higher and heating setpoint is lower in the night, which is possibly explained as for saving energy. This pattern is significant because any unfitted other pattern would be due to either the HVAC system malfunctioning in the corresponding zone or the setpoints being changed deliberately. We also find that a different but consistency setting on Day8 and Day 13 in every HVAC zone but Floor 3 Zone 10. The chart below shows the linked pattern (both normal unchanged and tweaked on Day 8 and Day 13) of thermostat temperature and two setpoints.

The pie charts of outliers regarding the thermostat temperature and setpoints also reflects a pattern of outliers. They indicate the outliers show up in Day 7 and Day 8 in most zones.

MC2.3 – Describe up to ten notable anomalies or unusual events you see in the data. Describe when and where the event or anomaly occurs and describe why it is notable. If you have more than ten anomalies to report, prioritize those anomalies that are most likely to represent a danger or serious issue for building operation.

Limit your response to no more than 10 images and 1000 words.

1. For the building data, most of the CO2 sensors start to detect unusual readings from Day 7, the situation last for the whole day and only alleviate after 10PM, and getting better after midnight, but it deteriorates when office hours starts. The outlier summary pie chart on the right also indicates that Day 7 and Day 8 have most outliers count during the 14 days.

2. During June 7 and June 8, around 7:00AM-10:00PM, the VAV system outdoor air flow fraction was set to maximum due to majority of the CO2 concentration readings reached a very high level.

3. In order to keep the HVAC running so as to “wash out” the excessive CO2, the thermostat Heating setpoint and Cooling setpoint were set to a very low lever, around 12/15 Celsius degree respectively. This cause a very low thermostat temperature before working hours. As the following image depicts, take HVAC F_2_Z_3 as an example, Day 7 and Day 8 setpoints (the lighter blue/red lines) as well as thermostat temperature are significantly lower than the other day. Most of the zones have this unusual event happened in Day 7 and Day 8.

4. The peak at 7:05 AM for Day 7 and Day 8 in reheat coil power and supply inlet temperature are the reaction of previous setpoint settings. After a night of HVAC running, it was too cold for the personnel starting to work in the morning. Therefore, they set both setpoint to a very high temperature until off work, and thus to maintain a comfortable thermostat temperature, the reheat power peaked 14-day high at that time for most of the zones. The supply inlet temperature also reached around 40 Celsius degrees.

5. While the thermostat setpoints and the corresponding thermostat temperature are being tweaked in all other zones on Day 7 and Day 8, nothing is found being changed in HVAC Floor 3 Zone 1. However, the thermostat setpoints are being lowered extensively on Day 13 around before dawn. And the reheat coil power is lowered much as well.

6. In terms of the employees’ abnormality, Geneviere Florez often work from 7:50 to 17:00. From the chart below we know that he is extremely punctual. But on June 7 he went to company in the middle of night and stay at PROX floor 1 zone 1 the whole night. Then went back to his office at normal hour.

7. Some employees have multiple id cards but they never use them at the same time. While Patrick young (pyoung001, pyoung002) use both of them on June 2 10:25:00 to enter floor 3 zone 1 which indicates two people were using these cards.

Besides, we find there is information confliction between his mobile data and procfixed data. Robot found pyoung001 on 06-03, 06-06, 06-07, 06-08, 06-09, 06-10, 06-13 morning around 9:10 to 9:25 am. While proximity zone sensor never detected he had entered floor 2 zone 6.

8. Sten Sanjorge Jr. went to company in the midnight of June 2, and stayed at floor 1 zone 1 until 2 PM. Then, he went back to his office.

9. Jeanetts Frost (Jfrost001) always stay at his own office. But on June 1, he went to floor 2 zone 2 at 10 AM and left at 2 PM.

10. Unlike other employees, Antonia Pinckney (apinckney001) and Ada Campo-Corrente (acampo001) rarely stay in their own offices.

MC2.4 –– Describe up to five observed relationships between the proximity card data and building data elements. If you find a causal relationship (for example, a building event or condition leading to personnel behavior changes or personnel activity leading to building operations changes), describe your discovered cause and effect, the evidence you found to support it, and your level of confidence in your assessment of the relationship.

Limit your response to no more than 10 images and 1000 words.

1. As we mentioned in answer 2.1.3, Room 103 the conference room (HVAC F_1_Z_5), corresponding to PROX Zone F_1_Z_6, held three meetings during weekday. This is a room in the middle of the building and not having natural lighting. Therefore, in this enclosed space, we know for sure, when a meeting is held, the light and the equipment is ON as the following image shows. We also noticed a latency after the people left. This phenomenon suggests it is possible that the light and equipment are automatically delayed off.

2. For the DELI on the floor 1 (HVAC F_1_Z_1, PROX F_1_Z_2), the zone population, thermostat temperature, and the CO2 concentration reach a peak during lunch time. The green line is 14-day median. We also noticed that the three peaks have chronological order. First the personnel enter the zone, and then the temperature rises due to activities, and finally these result in the CO2 rising.