Engagers Sense-Making Journal
====================

Mari, Marc, Çağsun, Antonio
MDEF: Measuring the world / A world in data activity report.

**Journal Index**

[TOC]

## From objectives to the hypothesis

### Brainstorming
We ideated on multiple directions we can take as a team and oriented the conversation towards ways in which we can measure class engagement.


### Project Goals
**Objective:**
We want to provide IAAC with schedule optimization insights for MDEF.

**Question:**
Does the level of student engagement change during a lecture?

**Hypothesis:**
Students are not actively engaged in the classroom and get distracted after a period of time.

### Reflection

Explain one or more mistakes you've done during that phase? What would you change if will do it again? (max 560 char)

There are several ways we could have improved our project. Firstly, we did not consider that the class' seating organization impedes some students from looking at the instructor at times. Some students metioned it was dificult to keep an eye on the LED when they had their backs to the instructors.

We also did not consider the class had two instructors, which meant that you could be paying attention to lecturer A while lecturer B is the one lighting up. This means that engaging with the exercise would induce disengagement from the class.

The LED lights were also extremeley bright, meaning those closer to the instructor would engage from the sudden change in light rather than class engagement itself.

Additionally, it must be considered different periods displayed different amounts of LED fades. The lesser the data, the lesser the accuracy levels for comparison.

## From hypothesis to data

### Tools selection

PHYSICAL INTERVENTION

TOOLS

Tools used for our Physical Intervention included: An Arduino ESP Featherboard, LED Lights, Broad Network, print-out data collection, an online survey and a classrom setting (including a teacher and engagers).

### Tool usage documentation

This physical intervention can be repeated continuously as long as there is a network running in a classroom setting. The process can be repated by launching the pre-established code through a network, and observing how students engage with the LED lights placed on the lecturer.

---

1. Randomized color Arduino Code

:::spoiler Project code can be accessed here

### Code

#include <WiFi.h>
#include <WiFiClient.h>
#include <WebServer.h>

const char* server = "maker.ifttt.com";

const char* resource = "/trigger/engagement_trigger/with/key/nnmdw9RaUq3b36e8fq4dhzIk_9cfEQGhKHUuOF5AldP";

const char* ssid = "...";
const char* password = "...";

int H1 = 0;
int H2 = 0;
int H3 = 0;
int H4 = 0;


#include <Adafruit_NeoPixel.h>
#ifdef __AVR__
#include <avr/power.h>
#endif
#define PIN 12
#define NUMPIXELS 60

Adafruit_NeoPixel pixels(NUMPIXELS, PIN, NEO_GRB + NEO_KHZ800);

uint32_t red = pixels.Color(255, 0, 0);
uint32_t blue = pixels.Color(0, 0, 255);
uint32_t green = pixels.Color(0, 255, 0);
uint32_t yellow = pixels.Color(255, 255, 0);
uint32_t color = red;

bool counting = false;
uint32_t timmer = 0;
uint32_t start = 0;



void setup() {
  pinMode(13, OUTPUT);
  digitalWrite(13,HIGH);
  pixels.begin();
  pixels.setBrightness(50);
  // Serial.begin(9600);
  Serial.begin(115200);  // PREGUNTARLE A JOSEP CUAL SE QUEDA O QP2
  Serial.println();
  Serial.println("Booting Sketch...");

  /*
//ESP32 As access point
  WiFi.mode(WIFI_AP); //Access Point mode
  WiFi.softAP(ssid, password);
*/
  //ESP32 connects to your wifi -----------------------------------
  WiFi.mode(WIFI_STA);  //Connectto your wifi
  WiFi.begin(ssid, password);

  Serial.println("Connecting to ");
  Serial.print(ssid);

  //Wait for WiFi to connect
  while (WiFi.waitForConnectResult() != WL_CONNECTED) {
    Serial.print(".");
  }

  //If connection successful show IP address in serial monitor
  Serial.println("");
  Serial.print("Connected to ");
  Serial.println(ssid);
  Serial.print("IP address: ");
  Serial.println(WiFi.localIP());  //IP address assigned to your ESP
  //----------------------------------------------------------------
}

void loop() {

  if (millis() < (60 * 60 * 1000 * 1)) {
    color = red;
  } else if (millis() < (60 * 60 * 1000 * 2)) {
    color = blue;
  } else if (millis() < (60 * 60 * 1000 * 3)) {
    color = green;
  } else if (millis() < (60 * 60 * 1000 * 4)) {
    color = yellow;
  }

  if (millis() - start > timmer) {
    Serial.println("changing color");
    pixels.fill(color, 0, 60);
    pixels.show();
    if (millis() < (60* 60 * 1000 * 1)) {
      H1 = H1 + 1;
    } else if (millis() < (60*60 * 1000 * 2)) {
      H2 = H2 + 1;
    } else if (millis() < (60 * 60 * 1000 * 3)) {
      H3 = H3 + 1;
    } else if (millis() < (60 * 60 * 1000 * 4)) {
      H4 = H4 + 1;
    }

    delay(5000);
    pixels.clear();
    pixels.show();
    counting = false;
    delay(1000);
    makeIFTTTRequest();
  }

  if (!counting) {
    timmer = random(5, 20) * 60000;
    Serial.println(timmer);
    counting = true;
    start = millis();
    Serial.print("Start:  ");
    Serial.println(start);
  }

  delay(5000);
}

// Make an HTTP request to the IFTTT web service
void makeIFTTTRequest() {
  
  WiFiClient client;
  int retries = 5;
  while(!!!client.connect(server, 80) && (retries-- > 0)) {
    Serial.print(".");
  }
  Serial.println();
  if(!!!client.connected()) {
    Serial.println("Failed to connect...");
  }
  

  // Temperature in Celsius
  String jsonObject = String("{\"value1\":\"") + H1 + "\",\"value2\":\"" + (H2)
                      + "\",\"value3\":\"" + H3 + "\"}";
                      
  // Comment the previous line and uncomment the next line to publish temperature readings in Fahrenheit                    
  /*String jsonObject = String("{\"value1\":\"") + (1.8 * bme.readTemperature() + 32) + "\",\"value2\":\"" 
                      + (bme.readPressure()/100.0F) + "\",\"value3\":\"" + bme.readHumidity() + "\"}";*/
                      
  client.println(String("POST ") + resource + " HTTP/1.1");
  client.println(String("Host: ") + server); 
  client.println("Connection: close\r\nContent-Type: application/json");
  client.print("Content-Length: ");
  client.println(jsonObject.length());
  client.println();
  client.println(jsonObject);
        
  int timeout = 5 * 10; // 5 seconds             
  while(!!!client.available() && (timeout-- > 0)){
    delay(100);
  }
  if(!!!client.available()) {
    Serial.println("No response...");
  }
  while(client.available()){
    Serial.write(client.read());
  }
  client.stop(); 
}

:::
<br>

2. LED strip changing color


3. LED strip necklace

4. Print-out Survey

5. Network color time and sequence Storage (IFTTT, Google Sheets)

6. Output print-out data

7. Online Survey

https://forms.gle/DnM1PrGVUNX9m3JH6

8. Excel Data Collection

9. Excel statistic Graphs

#### Data capturing strategy

How do you combine the tool provided with your creativity to prove your hypothesis? How long did you capture data?

Our approach to measure engagement involved a visual and manual method for data collection. Students in class were asked assist on data collection on the spot. Additionally, we configured a webhook in our code so that data would be posted and loged in a google sheets document.




#### Materials We Used

* ESP32 Featherboard Arduino
* LED String Light
* Clip (to clip apparatus to lecturer)
* Print-out Survey
* Subjects (Lecturer and Engagers)

#### Detail setup instructions

1. We ran the Arduino Code connected to the Wifi network, and started the proces
2. Fitted the LED strip necklace on the lecturer
3. Handed out the Print-out Survey to Students

4. Collected the print-out data


5. Imported data to excel manually and made visualisations of the results

#### Data collected

Class Start Time: **11:25**

Class End Time: **14:05**

Describe the raw data you collected by posting a sample i.e. a picture, a screen capture, etc.

#### ATTENDANCE ON THE DAY OF INTERVENTION


#### TABLE OF ACTUAL SIGNAL TIMESTAMPS

| 11:25 START | Red | Blue | Green |
|-|-|-|-|
| 11:40 | 1st | | |
| 11:47 | 2nd | | |
| 12:04 | 3rd | | |
| 12:17 | 4th | | |
| 12:26 | | 1st | |
| break 12:27 | | | |
| break 12:50 | | | |
| 13:09 | | 2nd | |
| 13:23 | | | 1st |
| 13:40 | | | 2nd |
| 13:51 | | | 3rd |
| 14:01 | | | 4th |
| 14:06 END | | | |

#### TALLY OF EACH STUDENT AGAINST ACTUAL SIGNALS

| | Period 1 (Red) | Period 2 (Blue) | Period 3 (Green) | Period 4 (Yellow) |
|-|-|-|-|-|
| Actual Value | 4 | 2 | 4 | 0 |
| Student # | | | | |
| 1 | 0 | 0 | 0 | 0 |
| 2 | 0 | 0 | 0 | 0 |
| 3 | 1 | 0 | 1 | 0 |
| 4 | 1 | 0 | 1 | 0 |
| 5 | 1 | 0 | 2 | 0 |
| 6 | 3 | 0 | 0 | 0 |
| 7 | 1 | 1 | 1 | 0 |
| 8 | 1 | 0 | 2 | 0 |
| 9 | 3 | 0 | 1 | 0 |
| 10 | 2 | 0 | 3 | 0 |
| 11 | 1 | 3 | 2 | 0 |
| 12 | 2 | 0 | 4 | 0 |
| 13 | 2 | 2 | 2 | 0 |
| 14 | 2 | 1 | 4 | 0 |
| 15 | 1 | 2 | 4 | 0 |
| 16 | 2 | 1 | 3 | 1 |
| 17 | 3 | 1 | 4 | 0 |
| 18 | 2 | 2 | 4 | 0 |
| 19 | 4 | 2 | 4 | 0 |
| 20 | 4 | 2 | 4 | 0 |

_*note that the color yellow was not displayed._

#### AVERAGE ACCURACY OF TALLIES ON EACH SECTION OF THE CLASS

Red: 1st Hour
Blue: 2nd Hour
Green: 3rd Hour (after break)


#### ACCURACY DISTRIBUTION OF THE CLASS (Each Column is a student)


### Learnings & Tips

- Obviously, measuring the light signals from the instructor is a proxy of the real engagement of the students. Looking at the instructor may not mean you're really engaged, and vice versa, there may be some neurodivergent people who can pay attention without looking at the instructor.

- It would be good to combine the activity mode of class (theory, hands-on, group work...) against the engagement level of the students.

- It would be ideal to do this type of study over a longer period of time with more diverse ways of collecting data to get a better understanding of the actual engagement in class.

- Self reporting survey at the end of this study didn't really work, as noone filled in the survey at the end of the session. It's a dilemma because we believe engagement can best be self-reported but asking people to get engaged with an additional element when you're studying their engagement may create bias in the study.

- Researching the engagement data of students from external resources mostly originate from engagement data of online learning sessions. We've reflected that measuring engagement in physical world is so much harder than in the digital space, which is evident in the way digital tools are obsessed about capturing our data as we use them.

## Data captured

### Data summary

| Data Summary | |
|--------------------------|---------------------------|
| Project Title | Mdef Engagement |
| Capture Start | 9-03-2023 |
| Capture End | 9-03-2023 |
| Original Data Format | Multiple images |
| Submitted format | CSV file |
| Total Data Points | 20 |
| Number of datasets | 1 |
| Data Repository | https://drive.google.com/file/d/1iPhQD3kSjyCDpgKvY3Ua29vpMweIm6_k/ |

## Data insights

Since students pay more attention at the start and and the end of the class, we suggest shorter time periods and a scheduled break in between.

**That is, instead of:**
10:00 hrs - 14:00 hrs

**We propose:**
10:00 hrs - 11:45 hrs Class pt.1
11:45 hrs - 12:15 hrs BREAK
12:15 hrs - 14:00 hrs Class pt.2

This would reduce the phase of class where students disconnect and begin to interact with other activities or disengage completely. We also consider that this design could make information easier to digest and also easier for instructors to organize and communicate.

---

## References

* https://randomnerdtutorials.com/esp32-esp8266-publish-sensor-readings-to-google-sheets/