Update README.md with correct instructions and links.

Author: evgkanias

README.md

@@ -6,32 +6,61 @@ Code that creates the analyses the data and figures of the article:
 *Celestial compass design mimics the fan-like polarisation filter array of insect eyes***.
 *Nature Communications*.
 
+## Clone the repository
+Open a terminal, clone the project, and navigate to its working directory:
+```commandline
+git clone https://github.com/InsectRobotics/CelestialCompassAnalysis.git
+cd CelestialCompassAnalysis
+```
+The processed data can be downloaded from [here](https://doi.org/10.7488/ds/6106) and
+they should be placed in the [csv](csv) directory. Alternatively, they can be generated by following the instructions below.
 
 ## Analyse the data and create the processed datasets
 
-The processed data can be downloaded from [here](). Alternatively, the data can be created
-by copying the ROS-bag files in the repositories workspace and following the instructions
-below.
+If you have access to the ROS-bag files produced by the robot (upon request from the authors),
+proceed with step 1a, otherwise proceed with step 1b.
 
-Open a terminal, navigate to the [templates](templates) directory, and run:
+### 1a. Generate the raw_dataset.csv
+Copy the ```sardinia_data``` and ```south_africa_data``` directories
+to the working directory, navigate to the [templates](templates) directory (e.g., ```cd templates```), and run the
+script that creates the raw dataset:
 ```commandline
 python create_csv.py -t raw
 ```
-This will create a CSV file in the ```csv``` directory (named ```raw_dataset.csv```) that
-contains all the important data from the ```sardinia_data``` and ```south_africa_data```
-directories (that contain the raw ROS-bag files). The size of these directories was almost 100 GB,
-so they were excluded from the dataset, but we include the code for reference. Providing the
-```raw_dataset.csv``` as input, should allow the following commands to work without problems.
+The ROS-bag files contain a lot of information, including low resolution videos captured by the fish-eye camera
+during the experiments that increased the size of the datasets to almost 100 GB. The above script creates a CSV
+file (named ```raw_dataset.csv```) in the [csv](csv) directory, which includes only the data used in the article
+without any further processing and reduces the size of the datasets to 677.7 MB. It also creates a directory in
+named ```sessions``` in the [csv](csv) directory, which contains the high-resolution fish-eye images captured at
+beginning of each session (400.1 MB).
+
+### 1b. Download the raw_dataset.csv
+Navigate to the [csv](csv) directory and download the raw dataset from the above link. Alternatively, run
+the below lines from the working directory: 
+```commandline
+cd csv
+wget https://datashare.ed.ac.uk/bitstream/handle/10283/7116/raw_dataset.csv?sequence=9&isAllowed=n
+wget https://datashare.ed.ac.uk/bitstream/handle/10283/7116/sessions.zip?sequence=5&isAllowed=n
+cs ..
+```
+This will download the CSV file (```raw_dataset.csv```) and fish-eye images (```sessions.zip```)
+in the ```csv``` directory. These contain all the important data from the ```sardinia_data``` and
+```south_africa_data``` directories (which are excluded from the dataset because of their large size).
+Finally, extract the ```sessions.zip``` in the [csv](csv) directory.
+
+### 2. Generate the pooled_dataset.csv
+Providing the ```raw_dataset.csv``` as input, should allow the following commands to work without problems.
 The option ```-o [output_path]``` allows for a different output path.
 To create the pooled data from the above CSV file, run:
 ```commandline
 python create_csv.py -t pooled -i DATASET_DIR/raw_dataset.csv
 ```
 which will create another CSV file in the ```csv``` directory (named ```pooled_dataset.csv```).
-Again the output path can be changed using the ```-o [output_path]``` option. This process
-needs the ```csv/raw_dataset.csv```, so if an alternative output path was chosen before,
-this has to be specified through the input option ```-i [input_path]```. Finally, the errors
-for all the sessions and models can be calculated by running:
+The ```-i DATASET_DIR/raw_dataset.csv``` part is optional, and if you followed step 1a or 1b, it
+shouldn't be necessary.
+
+### 3. Generate the error_dataset.csv
+The errors for all the sessions and models can be calculated by running:
 ```commandline
 python create_csv.py -t error
 ```