import pandas as pd
from matplotlib import pyplot as plt
%matplotlib inline

all_quakes = pd.DataFrame.from_csv("../data/earthquake_states.csv", index_col=None, parse_dates=["time", "updated"])

us_quakes = all_quakes.dropna(subset=["state"])

state_counts = pd.DataFrame(us_quakes.state.value_counts())

state_counts.columns = ["total_number_of_earthquakes"]

state_counts

ax = us_quakes[us_quakes["state"] == "Oklahoma"].set_index("time")["id"].resample("A", how="count").plot()
ax.set_title("Oklahoma Earthquake Count By Year")
pass

ax = us_quakes[us_quakes["state"] == "California"].set_index("time")["id"].resample("A", how="count").plot()
ax.set_title("California Earthquake Count By Year")
pass

ax = us_quakes[us_quakes["state"] == "Texas"].set_index("time")["id"].resample("A", how="count").plot()
ax.set_title("Texas Earthquake Count By Year")
pass

ax = us_quakes[us_quakes["state"] == "Ohio"].set_index("time")["id"].resample("A", how="count").plot(color="b")
ax.set_title("Ohio Earthquake Count By Year")
ax.set_ylim([0,5])
pass

ax = us_quakes[us_quakes["state"] == "Colorado"].set_index("time")["id"].resample("A", how="count").plot()
ax.set_title("Colorado Earthquake Count By Year")
ax.set_ylim([0,25])
pass

ax = us_quakes[us_quakes["state"] == "Tennessee"].set_index("time")["id"].resample("A", how="count").plot()
ax.set_title("Tennessee Earthquake Count By Year")
ax.set_ylim([0,10])
pass

ax = us_quakes[us_quakes["state"] == "Kentucky"].set_index("time")["id"].resample("A", how="count").plot()
ax.set_title("Kentucky Earthquake Count By Year")
ax.set_ylim([0,10])
pass

ax = us_quakes[us_quakes["state"] == "Kansas"].set_index("time")["id"].resample("A", how="count").plot()
ax.set_title("Kansas Earthquake Count By Year")
pass

ax = us_quakes[us_quakes["state"] == "Arkansas"].set_index("time")["id"].resample("A", how="count").plot()
ax.set_title("Arkansas Earthquake Count By Year")
pass

def quake_percentage_change(state):
    by_year = pd.DataFrame(us_quakes[us_quakes["state"] == state].set_index("time")["id"].resample("AS", how="count"))
    by_year["start"] = by_year.index
    by_year["year"] = by_year["start"].apply(lambda x: x.year)
    decade_05_14 = by_year[(by_year["year"] >= 2005) & (by_year["year"] <= 2014)]
    total_05_14 = decade_05_14["id"].sum()
    decade_95_04 = by_year[(by_year["year"] >= 1995) & (by_year["year"] <= 2004)]
    total_95_04 = decade_95_04["id"].sum()
    if total_95_04 != 0:
        pct = round(100.0 * (total_05_14 - total_95_04) / total_95_04, 2)
    else:
        pct = None
    return pct, total_05_14, total_95_04

state_counts["name"] = state_counts.index

state_counts["percentage_change"], state_counts["total_05-14"], state_counts["total_95-04"] =\
    zip(*state_counts["name"].apply(lambda x: quake_percentage_change(x)))

round(100.0 * (state_counts["total_05-14"].sum() - state_counts["total_95-04"].sum()) / state_counts["total_95-04"].sum(), 2)

state_counts[state_counts["total_95-04"] >= 5].sort("percentage_change", ascending=False)