// Load a file locally
viewof text = html`<input type=file * ">`
//turn JSON into text
tweetsStrings = Files.text(text)
//make text into dataframe
tweetsDF = Object.values(JSON.parse(tweetsStrings))
//create temporary file where we see what tweetsDF was originally
tweetsDF_temp = Object.values(JSON.parse(tweetsStrings))
//get unique location names from this raw data set
unique_locations_tweetsDF = z.unique(z.getCol("location",tweetsDF_temp)) //get unique location names
//group the raw tweets by user location
count_tweetsDF = {
let countArray = []
let groupedLocations = z.groupBy(x => x.location, tweetsDF_temp)
for (const key of Object.keys(groupedLocations)) {
countArray.push({location: key, count: groupedLocations[key].length})
}
return countArray
}
//sort by count of tweets; we see that variations of 'Boston, MA' are most common
countSorted_tweetsDF = z.sortByCol('count', 'des', count_tweetsDF)
// Loop through all the locations, and see if they are similar to Boston, then create a new column
newTweets = {
let cleanTweets = []
let locations = z.getCol("location", tweetsDF)
//for each row in tweetsDF...
for (let i in tweetsDF) {
//make a duplicate tweet structure to load into cleanTweets
let tweet = tweetsDF[i]
if (tweetsDF[i].location.toLowerCase().includes('boston')){
tweet.location = 'Boston, MA'
}
else if(tweetsDF[i].location.toLowerCase().includes('cambridge')){
tweet.location = 'Cambridge, MA'
}
else if(tweetsDF[i].location==""){
tweet.location = 'no location'
}
else {
//push the clean tweet to cleanTweets
}
cleanTweets.push(tweet)
}
return cleanTweets
}
//group newTweets by count
count_cleaned = {
let countArray = []
let groupedLocations = z.groupBy(x => x.location, newTweets)
for (const key of Object.keys(groupedLocations)) {
countArray.push({location: key, count: groupedLocations[key].length})
}
return countArray
}
//sort by count of tweets
countSorted_clean = z.sortByCol('count', 'des', count_cleaned)
vegalite({
data: {values: countSorted_clean.slice(0,10)},
mark: "bar",
encoding: {
x: {field: "location", type: "nominal", sort:"*"},
y: {field: "count", type: "quantitative"},
color: {"field": "location", "type": "nominal"}
}
})
//these aren't necessarily the stated locations of the users; they can be different from the "location"
notNull = z.filter(r => r.lat != null, tweetsDF)
vegalite({
data: {values: notNull},
mark: "point",
encoding: {
x: {field: "lat", type: "quantitative","scale": {"domain": [42.33,42.38]}},
y: {field: "lon", type: "quantitative","scale": {"domain": [-71,-71.15]}},
//color: {"field": "location", "type": "nominal"}
}
})
// Load a file locally
viewof hashtag_text = html`<input type=file * ">`
//turn JSON into text
tweetsStringsBrexit = Files.text(hashtag_text)
tweets_brexit = Object.values(JSON.parse(tweetsStringsBrexit))
//reorganizing tweets_brexit into the same structure as the first exercise
tweets_brexit_DF = {
let otherArray = []
for(let i in tweets_brexit){
let curr = tweets_brexit[i][1]
otherArray.push(curr)
}
return otherArray
}
//get all the locations in the tweets_brexit_DF
locationsArray = {
let newArray = []
for(let i in tweets_brexit_DF){
let currentText = tweets_brexit_DF[i].location
newArray.push(currentText)
}
return newArray
}
//get unique location names, there are a lot of England or cities in England
unique_locations_brexit = z.unique(z.getCol("location", tweets_brexit_DF))
//group by user location
count_brexit = {
let countArray = []
let groupedLocations = z.groupBy(x => x.location, tweets_brexit_DF)
for (const key of Object.keys(groupedLocations)) {
countArray.push({location: key, count: groupedLocations[key].length})
}
return countArray
}
//sort by count; lots of null locations
//locations have varying degrees of specificity (i.e. England; London, England; UK)
//chose to format London since it had a large number of entries as "London" or "London, England"
countSorted_brexit = z.sortByCol('count', 'des', count_brexit)
cleanedTweetsBrexit = {
let newLocations = []
let locations = z.getCol("location", tweets_brexit_DF)
for (let id in locations) {
if (locations[id].toLowerCase().includes('london')){
newLocations.push('London, England')
}
else if(locations[id]==""){
newLocations.push('null value')
}
else {
newLocations.push(locations[id])
}
}
return z.addCol("cleanLocations", newLocations, tweets_brexit_DF)
}
//none of the tweets collected have lat/lon
notNullBrexit = z.filter(r => r.lat != null, tweets_brexit_DF)
vegalite({
data: {values: countSorted_brexit},
mark: "bar",
encoding: {
x: {field: "lat", type: "quantitative"},
y: {field: "lon", type: "quantitative"},
color: {"field": "search", "type": "nominal"}
}
})
One platform to build and deploy the best data apps
Experiment and prototype by building visualizations in live JavaScript notebooks. Collaborate with your team and decide which concepts to build out.
Use Observable Framework to build data apps locally. Use data loaders to build in any language or library, including Python, SQL, and R.
Seamlessly deploy to Observable. Test before you ship, use automatic deploy-on-commit, and ensure your projects are always up-to-date.