import numpy as np import pandas as pd import matplotlib.pyplot as plt import seaborn as sns import plotly.graph_objects as go import plotly.express as px import os import warnings from sklearn.cluster import KMeans from sklearn.cluster import AgglomerativeClustering from sklearn.preprocessing import LabelEncoder from scipy.cluster import hierarchy from scipy.spatial import distance_matrix warnings.filterwarnings('ignore') os.listdir("data/") np.random.seed(2021)

df = pd.read_csv('./data/russian_alcohol_consumption.csv') df.head()

df.describe()

list(df['region'].unique())

dataplot = sns.heatmap(df.corr(), cmap="YlGnBu", annot=True) plt.show()

NA = pd.DataFrame( data=[ df.isna().sum().tolist(), [ "{:.2f}".format(i) + "%" for i in (df.isna().sum() / df.shape[0] * 100).tolist() ], ], columns=df.columns, index=["NA Count", "NA Percent"], ).T.sort_values(by="NA Count", ascending=False) NA.style.background_gradient(cmap="seismic", subset=["NA Count"])

df.fillna(method="pad", inplace=True) df.isna().sum()

geo = pd.read_csv("data/russian_geo.csv") df_geo = ( pd.merge(df, geo, on="region") .groupby("region") .mean() .reset_index() .sort_values("beer", ascending=False) )

df_geo["total"] = ( df_geo["wine"] + df_geo["brandy"] + df_geo["vodka"] + df_geo["beer"] + df_geo["champagne"] ) df_geo["text"] = ( df_geo["region"] + "<br> Alcohol consumption:" + (df_geo["total"]).astype(str) ) # limits = [(0,99),(100,299),(300,499),(500,999),(1000,3000)] limits = [(0, 10), (11, 21), (22, 84)] colors = ["royalblue", "crimson", "lightseagreen"] scale = 500 fig = go.Figure() for i in range(len(limits)): lim = limits[i] df_sub = df_geo[lim[0] : lim[1]] fig.add_trace( go.Scattergeo( lon=df_sub["lon"], lat=df_sub["lat"], text=df_sub["text"], marker=dict( size=df_sub["total"], color=colors[i], line_color="rgb(40,40,40)", line_width=0.5, sizemode="area", ), name="Top {0} - {1}".format(lim[0] + 1, lim[1]), ) ) fig.update_layout( title_text="Alcohol Consumption in Russia", showlegend=True, legend_title="Beer 1 liter/percapita", legend_title_font_size=14, geo=dict( scope="world", landcolor="rgb(217, 217, 217)", lonaxis=dict(range=[23.6985, 80.06269]), lataxis=dict(range=[33.72197, 80.7293]), ), ) fig.show()

fig, axes = plt.subplots(1, 5, figsize=(15, 6), sharey=True) for x in range(1, 6, 1): column_name = df.columns[x + 1] sns.lineplot(data=df, x="year", y=f"{column_name}", ax=axes[x - 1]) axes[x - 1].set_title(f"{column_name} consumption") axes[x - 1].set_xlabel("Year") axes[x - 1].set_ylabel("litres per capita")

year_df = ( df.groupby("year") .mean() .stack(0) .reset_index() .rename(columns={"level_1": "type", 0: "consumption"}) ) # year_df['total'] = year_df['wine']+year_df['brandy']+year_df['vodka']+year_df['beer']+year_df['champagne'] fig = px.scatter( year_df, x="type", y="consumption", animation_frame="year", color="type", size="consumption", size_max=55, # range_x=[100,100000], range_y=[0, 60], ) # fig["layout"].pop("updatemenus") # optional, drop animation buttons fig.show()

X = df.groupby(["region"])["wine", "beer", "vodka", "champagne", "brandy"].mean() clusters = [] for i in range(1, 10): km = KMeans(n_clusters=i).fit(X) clusters.append(km.inertia_) fig, ax = plt.subplots(figsize=(12, 8)) sns.lineplot(x=list(range(1, 10)), y=clusters, ax=ax) ax.set_title("Searching for Elbow") ax.set_xlabel("Clusters") ax.set_ylabel("Inertia");

km = KMeans(n_clusters=2).fit(X) X["Labels"] = km.predict(X) plt.figure(figsize=(12, 8)) sns.scatterplot(X["wine"], X["beer"], hue=X["Labels"], palette="Set1", alpha=0.7) plt.title("KMeans with 2 Clusters") plt.show()

km = KMeans(n_clusters=3).fit(X) X["Labels"] = km.predict(X) plt.figure(figsize=(12, 8)) sns.scatterplot(X["wine"], X["beer"], hue=X["Labels"], palette="Set1", alpha=0.7) plt.title("KMeans with 3 Clusters") plt.show()

fig, axes = plt.subplots(1, 4, figsize=(15, 6)) sns.scatterplot( X["wine"], X["beer"], hue=X["Labels"], palette="Set1", alpha=0.7, ax=axes[0] ) axes[0].set_title("beer V/S wine") sns.scatterplot( X["vodka"], X["champagne"], hue=X["Labels"], palette="Set1", alpha=0.7, ax=axes[1] ) axes[1].set_title("vodka V/S champagne") sns.scatterplot( X["wine"], X["champagne"], hue=X["Labels"], palette="Set1", alpha=0.7, ax=axes[2] ) axes[2].set_title("wine V/S champagne") sns.scatterplot( X["vodka"], X["brandy"], hue=X["Labels"], palette="Set1", alpha=0.7, ax=axes[3] ) axes[3].set_title("vodka V/S brandy") plt.show()

fig = plt.figure(figsize=(10, 8)) ax1 = fig.add_subplot(151) sns.swarmplot(x="Labels", y="beer", data=X, ax=ax1) ax1.set_title("beer") ax2 = fig.add_subplot(152) sns.swarmplot(x="Labels", y="vodka", data=X, ax=ax2) ax2.set_title("vodka") ax3 = fig.add_subplot(153) sns.swarmplot(x="Labels", y="wine", data=X, ax=ax3) ax3.set_title("wine") ax4 = fig.add_subplot(154) sns.swarmplot(x="Labels", y="champagne", data=X, ax=ax4) ax4.set_title("champagne") ax5 = fig.add_subplot(155) sns.swarmplot(x="Labels", y="brandy", data=X, ax=ax5) ax5.set_title("brandy") plt.show()

com_X = X["beer"] + X["wine"] + X["champagne"] + X["vodka"] + X["brandy"] sns.swarmplot(x="Labels", y=com_X, data=X) plt.title("Labels V/S All drinks");

dist = distance_matrix(X, X) Z = hierarchy.linkage(dist, "average")

plt.figure(figsize=(12, 8)) dendro = hierarchy.dendrogram(Z, leaf_rotation=0, leaf_font_size=6, orientation="right")

hc = AgglomerativeClustering(n_clusters=4, affinity="euclidean", linkage="ward") y_hc = hc.fit_predict(X)

plt.figure(figsize=(12, 8)) sns.scatterplot(X["wine"], X["beer"], hue=y_hc, palette="Set1", alpha=0.7) plt.title("Agglomerative Clustering with 4 Clusters") plt.show()

X["H_Pop"] = y_hc X["H_Pop"] = X["H_Pop"].replace(2, "Low") X["H_Pop"] = X["H_Pop"].replace(1, "Medium") X["H_Pop"] = X["H_Pop"].replace(3, "High") X["H_Pop"] = X["H_Pop"].replace(0, "Top") X[X["H_Pop"] == "Top"].head(2)

X.rename(columns={"Labels": "KM_Pop"}, inplace=True) X["Total_Drinks"] = com_X X["KM_Pop"] = X["KM_Pop"].replace(2, "Low") X["KM_Pop"] = X["KM_Pop"].replace(0, "Medium") X["KM_Pop"] = X["KM_Pop"].replace(1, "High") X[X["KM_Pop"] == "High"].sort_values("Total_Drinks", ascending=False).head(11)

city_names_total = ( X[(X["KM_Pop"] == "High") & (X["H_Pop"] == "Top")] .sort_values("Total_Drinks", ascending=False) .head(11) .index ) list(city_names_total.drop("Saint Petersburg"))

city_names_wine = ( X[(X["KM_Pop"] == "High") & (X["H_Pop"] == "Top")] .sort_values("wine", ascending=False) .head(11) .index ) list(city_names_wine.drop("Saint Petersburg"))

!nbqa black notebook.ipynb