Doing Data analytics on my Data
I created https://somethingto.do as a better way for me to interact with all the data that I was already sharing and storing in google maps. Part of the process to make the data easily accessible and parsable for me was by creating JSON files or JavaScript objects that would have all the important information for it to be useful and to function as a quasi-replacement of google maps.
Here is an example object:
{ "type": "Feature", "properties": { "name": "The New York Earth Room", "description": "I have been to this room many times and every time I find it amazing that they keep that dirt in it... it is pretty surreal to see because why would you do this? lol", "kind": "Gallery", "area": "SoHo, Manhattan", "visits": "3", "visited": true }, "geometry": { "type": "Point", "coordinates": [ 40.7260015, -73.9997858 ] } }
It has everything that I need for it to be useful to me. Name, Description or Review, Kind of place, Area, How many time I have been there or not, Have I visited this place? True or false, and geometric data (Latitude and longitude).
Inspiration
The inspiration to do this analysis is that I have fallen in love with data visualization and from the creation of something to do I have been wanting to create visualizations using D3.js and while have been able to do some of those interactive visualizations using D3, It has been difficult to do some other without running into issues.
The website that inspire me the most were some of the visualizations done by Jhonny Harris on his videos, The New York Times (The creator of D3 used to work at the times), and FiveThirtyEight.
Findings
I have been to a lot of places in New York City and it really felt that I have been to way more places all over the world than the 600 gone to as of March 2024.
I seem to love the neighborhoods south of 14ST and that is kind of true. I used to work in SoHo, and would visit a lot of cafes, restaurants, bars and more places in those neighborhoods.
This map on the other hand shows were most the places that I want to go to are. I want to travel more, since I get recommendations for places from all over the globe and I wanted to keep them in an somewhat accessible and easily to surface later to me.
The hot-spots are New York City (I live there), France (I have a lot of french friends), Italy (My french friends also love Italy or Sardinia), The Philippines (I have Filipino friends who gave me some reccs), and Spain.
My first Pie Chart visualization. This visualization shows how many categories and labels I used to have on my objects, this is way more categories and labels than I personally anticipated.
After seeing how huge and redundant it was I simplified it a little bit.
Second Iteration.
Third and last Iteration.
The simplified format will only live in this analysis as I actually liked the segmentation of restaurants 🙂
It makes it easier for me to know which kind of restaurant it is, and if I want to have different icons based on the type of restaurant later on it is also easier to implement.
Most Visited Area
My most visited area for all the categories is SoHo / East Village / LES.
Which makes sense I used to work in that area for couple of years, and spend a lot of time in that area.
Where have I been?
Look and you shall see!
Yes I have not traveled intensively yet
These are some of the areas that I would love to go too. There are a lot of parks and monuments represented on the map below, I enjoy history and nature what can say.
Favorite Kind of place
Restaurants. It make sense I eat everyday and a lot of the time I try to go to a restaurant to try something new.
Favorite kind of restaurant
I love Japanese food and that shows here. I usually separate Sushi, Ramen, and general Japanese restaurants but here I had then together to make easier to read.
How many places are there?
My list have 1068 places in total. That means places that I have been and I want to go.
Code
Bar Charts
import matplotlib.pyplot as plt import pandas as pd import json # Load the data with open('food.json') as f: data = json.load(f) # Convert the data to a DataFrame df = pd.json_normalize(data['features']) # Filter the DataFrame based on the 'visited' property df_visited = df[df['properties.visited'] == True] # Count the number of each type of place counts = df_visited['properties.kind'].value_counts() # Count the number of each type of place counts = counts.sort_values() # Create a bar chart plt.figure(figsize=(12, 16)) bars = plt.barh(counts.index, counts.values, height=0.5) # Add the total amount of places next to each bar for bar in bars: width = bar.get_width() plt.text(width, bar.get_y() + bar.get_height()/2, f' {width}', va='center') plt.ylabel('Type of Place') plt.xlabel('Number of Places') plt.title('Number of Each Type of Place') plt.xticks(rotation=90) plt.show()
Heat-maps of visited and not visited places
import matplotlib.pyplot as plt import folium import json from folium.plugins import HeatMap import pandas as pd with open('everything.json') as f: data = json.load(f) # Convert the list of dictionaries into a DataFrame df = pd.json_normalize(data['features']) df['latitude'] = df['geometry.coordinates'].apply(lambda x: x[0] if len(x) > 1 else None) df['longitude'] = df['geometry.coordinates'].apply(lambda x: x[1]if len(x) > 1 else None) # Filter the DataFrame based on the 'visited' attribute df_visited = df[df['properties.visited'] == True] df_not_visited = df[df['properties.visited'] == False] # Create a base map m = folium.Map([40.7128, -74.0060], zoom_start=11, tiles='CartoDB Positron') # Create a list of [latitude, longitude] pairs for visited and not visited places locations_visited = df_visited[['latitude', 'longitude']].values.tolist() locations_not_visited = df_not_visited[['latitude', 'longitude']].values.tolist() # Add a heatmap for visited places HeatMap(locations_visited).add_to(m) # Create a new map for not visited places m_not_visited = folium.Map([40.7128, -74.0060], zoom_start=11, tiles='CartoDB Positron') # Add a heatmap for not visited places HeatMap(locations_not_visited).add_to(m_not_visited) # Display the maps m.save('heatmap_visited.html') m_not_visited.save('heatmap_not_visited.html')
Pie chart
import json import matplotlib.pyplot as plt from matplotlib.patches import Patch # Load the data with open('everything.json') as f: data = json.load(f) # Extract the kinds of places and their counts kinds = {} for feature in data["features"]: kind = feature["properties"].get("kind", None) if kind is not None: if kind in kinds: kinds[kind] += 1 else: kinds[kind] = 1 # Calculate the total count total = sum(kinds.values()) # Sort the kinds by count in descending order kinds = dict(sorted(kinds.items(), key=lambda item: item[1], reverse=True)) # Create the pie chart plt.figure(figsize=(20, 10)) # Adjust the size as needed explode = [0.1 if v < sum(kinds.values()) * 0.05 else 0 for v in kinds.values()] wedges, _ = plt.pie(kinds.values(), labels=None, explode=explode) # Create the legend labels with percentages labels = [f'{k} - {v / total * 100:.1f}%' for k, v in kinds.items()] # Add the total count to the legend labels.append(f'Total - {total}') # Create legend handles handles = wedges + [Patch(facecolor='none')] legend = plt.legend(handles, labels, title="Kind of Place", bbox_to_anchor=(1, 0, 0, 1), loc="center left", ncol=2, fontsize='small') # Make the total count bold legendHandles = legend.legendHandles legendHandles[-1].set_visible(False) legend.texts[-1].set_weight('bold') plt.axis('equal') # Equal aspect ratio ensures that pie is drawn as a circle. # plt.tight_layout() # Manually adjust the subplot parameters plt.subplots_adjust(right=0.60) # Adjust this value as needed to make room for the legend plt.show()
Colored Heat-map
import json import pandas as pd import folium from folium.plugins import HeatMap from branca.element import Template, MacroElement # Load your GeoJSON data with open('everything.json', 'r') as f: geojson_data = json.load(f) # Convert the GeoJSON data to a DataFrame df = pd.json_normalize(geojson_data['features']) # Extract latitude and longitude from the 'geometry.coordinates' column df['latitude'] = df['geometry.coordinates'].apply(lambda x: x[0] if len(x) > 1 else None) df['longitude'] = df['geometry.coordinates'].apply(lambda x: x[1]if len(x) > 1 else None) # Define a color gradient for each type of place colors = { 'Cafe': {0.2: 'purple', 0.4: 'purple', 0.6: 'purple', 0.8: 'purple', 1: 'purple'}, 'Bar': {0.2: 'orange', 0.4: 'orange', 0.6: 'orange', 0.8: 'orange', 1: 'orange'}, 'Restaurant': {0.2: 'teal', 0.4: 'teal', 0.6: 'teal', 0.8: 'teal', 1: 'teal'}, 'Speakeasy': {0.2: 'salmon', 0.4: 'salmon', 0.6: 'salmon', 0.8: 'salmon', 1: 'salmon'}, 'Club': {0.2: 'yellow', 0.4: 'yellow', 0.6: 'yellow', 0.8: 'yellow', 1: 'yellow'}, 'Park': {0.2: 'lime', 0.4: 'lime', 0.6: 'lime', 0.8: 'lime', 1: 'lime'}, # I found that using the same color gives a more pleasent feeling and cohesiveness # to the heatmap # 'Cafe': {0.2: 'blue', 0.4: 'purple', 0.6: 'pink', 0.8: 'orange', 1: 'red'}, # 'Bar': {0.2: 'green', 0.4: 'yellow', 0.6: 'orange', 0.8: 'red', 1: 'maroon'}, # 'Restaurant': {0.2: 'navy', 0.4: 'blue', 0.6: 'aqua', 0.8: 'teal', 1: 'green'}, # 'Speakeasy': {0.2: 'maroon', 0.4: 'red', 0.6: 'salmon', 0.8: 'pink', 1: 'white'}, # 'Club': {0.2: 'black', 0.4: 'brown', 0.6: 'orange', 0.8: 'yellow', 1: 'white'}, # 'Park': {0.2: 'darkgreen', 0.4: 'green', 0.6: 'lime', 0.8: 'yellow', 1: 'white'}, # Add more types of places and colors as needed } # Create a base map m = folium.Map([40.7128, -74.0060], zoom_start=11, tiles='CartoDB Positron') # Create a separate heatmap for each type of place for place_kind, color in colors.items(): # Filter the DataFrame based on the type of place df_filtered = df[df['properties.kind'] == place_kind] # Add a heatmap to the base map HeatMap(data=df_filtered[['latitude', 'longitude']].dropna().groupby(['latitude', 'longitude']).sum().reset_index().values.tolist(), gradient=color, radius=8, max_zoom=13).add_to(m) # Create a legend and added to the HTML template = """ {% macro html(this, kwargs) %} <div style=" position: fixed; bottom: 100px; left: 50px; width: 200px; height: 110px; z-index:9999; font-size:14px; "> <p><a style="color:#A020F0;">█</a> Cafe</p> <p><a style="color:#FF7F00;">█</a> Bar</p> <p><a style="color:#008080;">█</a> Restaurant</p> <p><a style="color:#FA8072;">█</a> Speakeasy</p> <p><a style="color:#FFFF00;">█</a> Club</p> <p><a style="color:#00FF00;">█</a> Park</p> </div> {% endmacro %} """ macro = MacroElement() macro._template = Template(template) m.get_root().add_child(macro) m.save('heatmap.html')
If you would like to follow my journey here is the website!
