def GetLocalGig():
variables = {}
variables = load_variables()
response = couchdb_requests.couch_get_view(variables, "twitter/",
"_design/LocalGig/",
"_view/local-gig-view",
"?reduce=true&group_level=1")
statemap = {}
for row in response["rows"]:
statemap[row["key"][0]] = row["value"]
statemap = state_sorter_twitter(statemap)
statekeys = list(statemap.keys())
statevalues = list(statemap.values())
response = couchdb_requests.couch_get_view(variables, "twitter/",
"_design/LocalGig/",
"_view/local-gig-view",
"?reduce=true&group_level=2")
fig, ax = plt.subplots()
ax.bar(statekeys, statevalues)
ax.set_ylabel('Count')
ax.set_title(
'Num of Gig economy was tweeted in states of Australia from 2010 to 2020'
)
ax.set_xticklabels(statekeys)
for a, b in zip(statekeys, statevalues):
plt.text(a,
b + 130,
str(b),
horizontalalignment='center',
verticalalignment='center')
fig.savefig('img/stateGig.png')
return "img/stateGig.png"
def get_business_popularity():
variables = {}
variables = load_variables()
# Extract popular keyword view
response = couchdb_requests.couch_get_view(variables, "twitter/",
"_design/keyword_count/",
"_view/keyword_popular/",
"?reduce=true&group_level=1")
#store response in dictionary
keyword_count_dict = {}
for row in response["rows"]:
keyword_count_dict[row["key"]] = row["value"]
y_pos = list(keyword_count_dict.keys())
x_pos = list(keyword_count_dict.values())
#plot the graph
plt.rcdefaults()
fig, ax = plt.subplots(figsize=(10, 10))
y_arrange = np.arange(len(y_pos))
ax.barh(y_arrange, x_pos, align='center', color="red")
ax.set_yticks(y_arrange)
ax.set_yticklabels(y_pos)
ax.invert_yaxis() # labels read top-to-bottom
ax.set_xlabel('Number of Tweets')
ax.set_title('Popularity of Gig Economy Businesses')
#save and return image path
fig.savefig('img/keyword_pop.png')
return "img/keyword_pop.png"
def GetUsefulKeywords():
variables = {}
variables = load_variables()
response = couchdb_requests.couch_get_view(
variables, "twitter/", "_design/GigSentimental/",
"_view/sentimentOnStateKeywordYear", "?reduce=true&group_level=3")
usefulKeywords = []
keywordsList = [
'airbnb', 'airly', 'airtasker', 'bettercaring', 'camplify',
'carnextdoor', 'classbento', 'deliveroo', 'designcrowd', 'doordash',
'ebay', 'etsy', 'fiverr', 'gocatch', 'gumtree', 'helpling', 'homeaway',
'lyft', 'menulog', 'olacab', 'parkhound', 'pawshake', 'ratesetter',
'redbubble', 'shebah', 'sidekicker', 'spacer', 'stayz', 'stellar',
'uber', 'upwork', 'urbansitter', 'zomato', 'zoom2u'
]
locList = [
'Melbourne', 'Adelaide', 'Brisbane', 'Canberra', 'Darwin', 'Hobart',
'Perth', 'Sydney', 'Victoria', 'South Australia', 'Queensland',
'New South Wales', 'Northern Territory', 'Tasmania',
'Western Australia'
]
for i in response["rows"]:
location = i["key"][0]
word = i["key"][1]
year = i["key"][2]
# only want data for the particular keyword selected in 2017
if (word in keywordsList and year == "2017"):
if location in locList:
# check if the keyword already exists in the list or not
if (word not in usefulKeywords):
usefulKeywords.append(word)
return usefulKeywords
def get_income_tweet():
variables = {}
variables = load_variables()
# Get the Aurin Income Data
response = couchdb_requests.couch_get_view(variables, "aurin-mean-income/",
"_design/city-income/",
"_view/gccsavsincome-view/", "")
#Parse the response
city_income_dict = {}
for row in response["rows"]:
city_income_dict[row["key"]] = row["value"]
#Get the twitter data for that location
city_income_dict = state_sorter_aurin(city_income_dict)
response = couchdb_requests.couch_get_view(variables, "twitter/",
"_design/location-keyword/",
"_view/locationvskeyword/",
"?reduce=true&group_level=1")
#Parse the response
city_count_dict = {}
for row in response["rows"]:
city_count_dict[row["key"][0]] = row["value"]
city_count_dict = state_sorter_twitter(city_count_dict)
# create graph
incomes = list(city_income_dict.values())
tweet_num = list(city_count_dict.values())
states = list(city_count_dict.keys())
fig, ax = plt.subplots()
ax.scatter(incomes, tweet_num)
ax.set_xlabel('Mean Income ($)')
ax.set_ylabel('Number of Tweets')
for i in range(len(incomes)):
ax.annotate(
' ' + states[i],
(incomes[i], tweet_num[i]),
)
#save and return image
fig.savefig('img/incomevstweets.png')
return "img/incomevstweets.png"
def Getunemploy():
variables = {}
variables = load_variables()
response = couchdb_requests.couch_get_view(variables,
"aurin-employment-gcc/",
"_design/states_unemploy/",
"_view/unemploy_rate",
"?reduce=true&group_level=3")
unemployrate = defaultdict(float)
for i in response["rows"]:
location = i["key"][0]
total = i["key"][1] + i["key"][2] + i['value']
rate = i["value"] / total # calulate the unemployment rate
# group value by state
if location == "New South Wales":
unemployrate["NSW"] += round(rate, 4)
elif location == "Queensland":
unemployrate["QLD"] += round(rate, 4)
elif location == "South Australia":
unemployrate["SA"] += round(rate, 4)
elif location == "Victoria":
unemployrate["VIC"] += round(rate, 4)
elif location == "Western Australia":
unemployrate["WA"] += round(rate, 4)
# start plot
statekeys = list(unemployrate.keys())
statevalues = list(unemployrate.values())
fig, ax = plt.subplots()
ax.barh(statekeys, statevalues, height=0.5)
ax.set_xlabel('Unemployment Rate', fontsize=12)
ax.set_ylabel('States in Australia', fontsize=12)
ax.set_title('Unemployment Rate in 2018', fontsize=16)
ax.set_yticklabels(statekeys)
for a, b in zip(statevalues, statekeys):
plt.text(a - 0.003,
b,
'{:.2%}'.format(a),
horizontalalignment='center',
verticalalignment='center')
fig.savefig('img/unemployment_rate.png')
return 'img/unemployment_rate.png', unemployrate
def get_business_pop_location():
variables = {}
variables = load_variables()
# Extract popular keyword view
response = couchdb_requests.couch_get_view(variables, "twitter/",
"_design/location-keyword/",
"_view/locationvskeyword/",
"?reduce=true&group_level=2")
# Parse the response and store popularity according to each state
states = ['VIC', 'NSW', 'WA', 'SA', 'NT', 'ACT', 'QLD', 'TAS']
keyword_loc_dict = {}
for key in states:
keyword_loc_dict[key] = {}
for row in response["rows"]:
if (row["key"][0] == 'OT'):
pass
else:
keyword_loc_dict[row["key"][0]].update(
{row["key"][1]: row["value"]})
images = []
#create graph for each state
for state in states:
y_pos = list(keyword_loc_dict[state].keys())
x_pos = list(keyword_loc_dict[state].values())
plt.rcdefaults()
fig, ax = plt.subplots(figsize=(10, 10))
y_arrange = np.arange(len(y_pos))
ax.barh(y_arrange, x_pos, align='center')
ax.set_yticks(y_arrange)
ax.set_yticklabels(y_pos)
ax.invert_yaxis() # labels read top-to-bottom
ax.set_xlabel('Number of Tweets')
ax.set_title('Popularity of Gig Economy Businesses in ' + state)
#save figure for state
fig.savefig('img/keyword_loc_' + state + '.png')
images.append('img/keyword_loc_' + state + '.png')
return images
def get_unemployment_tweet():
variables = {}
variables = load_variables()
#get unemployment data from view
response = couchdb_requests.couch_get_view(variables,
"aurin-employment-sa2/",
"_design/unemployment_doc/",
"_view/sa2_decdata/",
"?reduce=true&group_level=1")
#parse the response
unemployment_count = {}
for row in response["rows"]:
unemployment_count[row["key"]] = row["value"]
unemployment_count['QLD'] = [
4348, 4386, 4950, 4528, 5129, 5872, 5941, 5664
]
total = []
#get tweets based on year
response = couchdb_requests.couch_get_view(variables, "twitter/",
"_design/location-keyword/",
"_view/year_keyword/",
"?reduce=true&group_level=1")
#parse response for graphing
year_count = {}
for row in response["rows"]:
year_count[row["key"][0]] = row["value"]
#get tweets based on location
response = couchdb_requests.couch_get_view(variables, "twitter/",
"_design/location-keyword/",
"_view/year_location/",
"?reduce=true&group_level=2")
year_location_count = {}
for key in unemployment_count.keys():
year_location_count[key] = {}
for row in response["rows"]:
year_location_count[row["key"][1]].update(
{row["key"][0]: row["value"]})
#create graph one: unemployment vs year per state
x_axis = ['2010', '2011', '2012', '2013', '2014', '2015', '2016', '2017']
plt_1 = plt.figure(1, figsize=(10, 10))
for state in unemployment_count.keys():
plt.plot(x_axis, unemployment_count[state], label=state)
plt.xlabel('year', fontsize=14)
plt.ylabel('no. of people unemployed', fontsize=14)
plt.title('Unemployment vs Year (State)', fontsize=20)
plt_1.legend(loc=7)
plt_1.savefig('img/unemp_vs_year_state.png')
#create graph two: unemployment vs year overall
total = np.zeros(len(unemployment_count['VIC']))
for value in unemployment_count.values():
for i in range(len(value)):
total[i] += value[i]
plt_2 = plt.figure(2, figsize=(8, 8))
plt.plot(x_axis, total)
plt.xlabel('year', fontsize=14)
plt.ylabel('no. of people unemployed', fontsize=14)
plt.title('Unemployment vs Year', fontsize=20)
plt_2.savefig('img/unemp_vs_year.png')
#create graph two: unemployment vs year overall
x_axis_4 = list(year_count.keys())
y_axis_4 = []
for year in x_axis:
y_axis_4.append(year_count[year])
plt_3 = plt.figure(3, figsize=(8, 8))
plt.plot(x_axis, y_axis_4, color="red")
plt.xlabel('year', fontsize=16)
plt.ylabel('no. of gig economy tweets', fontsize=16)
plt.title('Gig Economy Tweets vs year', fontsize=20)
plt_3.savefig('img/tweets_vs_year_state.png')
#save and return images
images = [
'img/unemp_vs_year_state.png', 'img/unemp_vs_year.png',
'img/tweets_vs_year_state.png'
]
return images
def GetKeywordSentiment(keyword):
variables = {}
variables = load_variables()
response = couchdb_requests.couch_get_view(
variables, "twitter/", "_design/GigSentimental/",
"_view/sentimentOnStateKeywordYear", "?reduce=true&group_level=3")
statemap = initialiseStateMap()
freq = defaultdict(int)
for i in response["rows"]:
location = i["key"][0]
word = i["key"][1]
year = i["key"][2]
# only want data for the particular keyword selected in 2017
# data processing
if (word == keyword and year == "2017"):
value = i["value"]
if location == "Melbourne":
statemap["VIC"] += value
freq["VIC"] += 1
elif location == "Adelaide":
statemap["SA"] += value
freq["SA"] += 1
elif location == "Brisbane":
statemap["QLD"] += value
freq["QLD"] += 1
elif location == "Canberra":
statemap["NSW"] += value
freq["NSW"] += 1
elif location == "Darwin":
statemap["NT"] += value
freq["NT"] += 1
elif location == "Hobart":
statemap["TAS"] += value
freq["TAS"] += 1
elif location == "Perth":
statemap["WA"] += value
freq["WA"] += 1
elif location == "Sydney":
statemap["NSW"] += value
freq["NSW"] += 1
elif location == "Victoria":
statemap["VIC"] += value
freq["VIC"] += 1
elif location == "South Australia":
statemap["SA"] += value
freq["SA"] += 1
elif location == "Queensland":
statemap["QLD"] += value
freq["QLD"] += 1
elif location == "New South Wales":
statemap["NSW"] += value
freq["NSW"] += 1
elif location == "Northern Territory":
statemap["NT"] += value
freq["NT"] += 1
elif location == "Tasmania":
statemap["TAS"] += value
freq["TAS"] += 1
elif location == "Western Australia":
statemap["WA"] += value
freq["WA"] += 1
else:
statemap[location] += value
freq[location] += 1
# get the avg (rounding to 2 decimal places)
for state, value in statemap.items():
if (freq[state] != 0):
value = (value / freq[state] * 100)
statemap[state] = round(value, 2)
# store the state with the lowest sentiment value and the sentiment value
(lowestSentValue, lowestSentState) = min(
(value, key) for key, value in statemap.items() if value != 0)
isHypothesisTrue = False
# assume negative sentiment initially
isTasNegative = True
isSANegative = True
# hypothesis true
if (statemap["TAS"] < 0 and statemap["SA"] < 0):
isHypothesisTrue = True
# check if tasmania has positive sentiment
if (statemap["TAS"] > 0):
isTasNegative = False
# check if south australia has positive sentiment
if (statemap["SA"] > 0):
isSANegative = False
statekeys = list(statemap.keys())
statevalues = list(statemap.values())
# bar chart plotting
fig, ax = plt.subplots()
# draw a line on y=0
plt.axhline(y=0, linestyle='-', color='black')
ax.bar(statekeys, statevalues, color='green')
ax.set_ylabel('Sentiment (%)')
ax.set_title('Sentiment Value on Gig Economy Based \n on \"' + keyword +
'\" keyword ' + 'for States of Australia',
loc='center')
ax.set_xticklabels(statekeys)
for a, b in zip(statekeys, statevalues):
if (b == 0):
plt.text(a, b, "", horizontalalignment='center')
elif (b < 0):
plt.text(a, b - 1.7, str(b), horizontalalignment='center')
else:
plt.text(a, b + 0.4, str(b), horizontalalignment='center')
fig.savefig('img/keywordSentiment.png')
return ("img/keywordSentiment.png", isHypothesisTrue, isTasNegative,
isSANegative, lowestSentState, lowestSentValue)
def GetElderlyPopPercentage():
variables = {}
variables = load_variables()
response = couchdb_requests.couch_get_view(variables, "aurin-population/",
"_design/population/",
"_view/population/", "")
statemap = defaultdict(int)
freq = defaultdict(int)
# going through the twitter data from couchdb
# do data processing
print(response)
for i in response["rows"]:
location = i["key"]
value = i["value"]["elderly_pop_pr100"]
if location == "Greater Melbourne":
statemap["VIC"] += value
freq["VIC"] += 1
elif location == "Greater Adelaide":
statemap["SA"] += value
freq["SA"] += 1
elif location == "Greater Brisbane":
statemap["QLD"] += value
freq["QLD"] += 1
elif location == "Greater Darwin":
statemap["NT"] += value
freq["NT"] += 1
elif location == "Greater Hobart":
statemap["TAS"] += value
freq["TAS"] += 1
elif location == "Greater Perth":
statemap["WA"] += value
freq["WA"] += 1
elif location == "Greater Sydney":
statemap["NSW"] += value
freq["NSW"] += 1
elif location == "Rest of Vic.":
statemap["VIC"] += value
freq["VIC"] += 1
elif location == "Rest of SA":
statemap["SA"] += value
freq["SA"] += 1
elif location == "Rest of Qld":
statemap["QLD"] += value
freq["QLD"] += 1
elif location == "Rest of NSW":
statemap["NSW"] += value
freq["NSW"] += 1
elif location == "Rest of NT":
statemap["NT"] += value
freq["NT"] += 1
elif location == "Rest of Tas.":
statemap["TAS"] += value
freq["TAS"] += 1
elif location == "Rest of WA":
statemap["WA"] += value
freq["WA"] += 1
else:
statemap[location] += value
freq[location] += 1
# get the avg (after rounding to 2 decimal places)
for state, value in statemap.items():
statemap[state] = round(value / freq[state], 1)
statekeys = list(statemap.keys())
statevalues = list(statemap.values())
# bar chart plotting
fig, ax = plt.subplots()
ax.bar(statekeys, statevalues, color='blue')
ax.set_ylabel('Population (%)')
ax.set_title('Elderly Population for States of Australia in 2017')
ax.set_xticklabels(statekeys)
for a, b in zip(statekeys, statevalues):
plt.text(a,
b + 0.5,
str(b),
horizontalalignment='center',
verticalalignment='center')
fig.savefig('img/elderlyAgePop.png')
return "img/elderlyAgePop.png"
def GetsentimentD():
variables = {}
variables = load_variables()
response = couchdb_requests.couch_get_view(variables, "twitter/",
"_design/regionVSsentiment/",
"_view/senti_date",
"?reduce=true&group_level=2")
count = defaultdict(int)
statemap = defaultdict(int)
for i in response["rows"]:
location = i["key"][0]
date = i["key"][1]
year = date.split(' ')[-1]
# find the data that in 2018
if year == '2018':
value = i["value"]['sum']
if location == "Melbourne":
count['VIC'] += 1
statemap["VIC"] += value
elif location == "Adelaide":
count['SA'] += 1
statemap["SA"] += value
elif location == "Brisbane":
count['QLD'] += 1
statemap["QLD"] += value
elif location == "Canberra":
count['NSW'] += 1
statemap["NSW"] += value
elif location == "Darwin":
count['NT'] += 1
statemap["NT"] += value
elif location == "Hobart":
count['TAS'] += 1
statemap["TAS"] += value
elif location == "Perth":
count['WA'] += 1
statemap["WA"] += value
elif location == "Sydney":
count['NSW'] += 1
statemap["NSW"] += value
elif location == "Victoria":
count['VIC'] += 1
statemap["VIC"] += value
elif location == "South Australia":
count['SA'] += 1
statemap["SA"] += value
elif location == "Queensland":
count['QLD'] += 1
statemap["QLD"] += value
elif location == "New South Wales":
statemap["NSW"] += value
elif location == "Northern Territory":
count['NT'] += 1
statemap["NT"] += value
elif location == "Tasmania":
count['TAS'] += 1
statemap["TAS"] += value
elif location == "Western Australia":
count['WA'] += 1
statemap["WA"] += value
else:
count[location] += 1
statemap[location] += value
# print(count)
for key in statemap.keys():
statemap[key] = statemap[key] / count[key]
statekeys = list(statemap.keys())
statevalues = list(statemap.values())
# start plot
fig, ax = plt.subplots()
ax.bar(statekeys, statevalues, width=0.5)
ax.set_ylabel('Sentiment Rate', fontsize=12)
ax.set_xlabel('States in Australia', fontsize=12)
ax.set_title('Sentiment to Gig Economy in 2018', fontsize=16)
ax.set_xticklabels(statekeys)
for a, b in zip(statekeys, statevalues):
plt.text(a,
b + 0.0025,
round(b, 3),
horizontalalignment='center',
verticalalignment='center')
fig.savefig('img/sentiment2018.png')
return 'img/sentiment2018.png', statemap
def Getsentiment():
variables = {}
variables = load_variables()
response = couchdb_requests.couch_get_view(variables, "twitter/",
"_design/regionVSsentiment/",
"_view/senti",
"?reduce=true&group_level=1")
statemap = defaultdict(int)
for i in response["rows"]:
location = i["key"][0]
# calulate the sentiment value for each region.
value = i["value"]['sum'] / i["value"]['count']
# group value by state
if location == "Melbourne":
statemap["VIC"] += value
elif location == "Adelaide":
statemap["SA"] += value
elif location == "Brisbane":
statemap["QLD"] += value
elif location == "Canberra":
statemap["NSW"] += value
elif location == "Darwin":
statemap["NT"] += value
elif location == "Hobart":
statemap["TAS"] += value
elif location == "Perth":
statemap["WA"] += value
elif location == "Sydney":
statemap["NSW"] += value
elif location == "Victoria":
statemap["VIC"] += value
elif location == "South Australia":
statemap["SA"] += value
elif location == "Queensland":
statemap["QLD"] += value
elif location == "New South Wales":
statemap["NSW"] += value
elif location == "Northern Territory":
statemap["NT"] += value
elif location == "Tasmania":
statemap["TAS"] += value
elif location == "Western Australia":
statemap["WA"] += value
else:
statemap[location] += value
statekeys = list(statemap.keys())
statevalues = list(statemap.values())
fig, ax = plt.subplots()
ax.bar(statekeys, statevalues, width=0.5)
ax.set_ylabel('Sentiment Rate')
ax.set_title('Sentiment About Gig Economy')
ax.set_xticklabels(statekeys)
for a, b in zip(statekeys, statevalues):
plt.text(a,
b + 0.01,
round(b, 3),
horizontalalignment='center',
verticalalignment='center')
fig.savefig('img/sentiment.png')
return 'img/sentiment.png'