gistfile1.txt

from babel.numbers import format_currency

# from gooeysite import wsgi
#
# assert wsgi

# import streamlit as gui
import plotly.graph_objects as go
import pandas as pd

gui.set_page_config(layout="wide")


df = pd.read_csv("/Users/Dev/Downloads/wid_all_data/WID_data_US.csv", sep=";")
meta = pd.read_csv("/Users/Dev/Downloads/wid_all_data/WID_metadata_US.csv", sep=";")
# gui.write(df)
# gui.write(meta)

parse_percentile = lambda x: tuple(map(float, x.strip("p").split("p")))

groups = dict(
    key_groups=[
        "p0p100", "p0p50", "p50p90", "p90p99", "p99p100", "p99.99p100",
        # All deciles (p0p10, p10p20, p20p30, …, p90p100)
        *["p{}p{}".format(i * 10, (i + 1) * 10) for i in range(10)],
    ],
    detailed_p_groups=[
        # All percentiles (p0p1, p1p2, …, p98p99, p99p100)
        *["p{}p{}".format(i, i + 1) for i in range(99)],
        # Tenths of a percentile in the top 1%
        "p99p99.1", "p99.1p99.2", "p99.2p99.3", "p99.3p99.4", "p99.4p99.5", "p99.5p99.6", "p99.6p99.7", "p99.7p99.8", "p99.8p99.9",
        #"p99.9p100",
        # Hundreds of a percentile in the top 0.1%
        "p99.9p99.91", "p99.91p99.92", "p99.92p99.93", "p99.93p99.94", "p99.94p99.95", "p99.95p99.96", "p99.96p99.97", "p99.97p99.98", "p99.98p99.99",
        #"p99.99p100",
        # Thousands of a percentile in the top 0.01%
        "p99.99p99.991", "p99.991p99.992", "p99.992p99.993", "p99.993p99.994", "p99.994p99.995", "p99.995p99.996", "p99.996p99.997", "p99.997p99.998", "p99.998p99.999",
        # 10 thousands of a percentile in the top 0.001%
        "p99.999p100",
    ],
    # (p0p100, p1p100, p2p100, …)
    detailed_top_groups=["p{}p100".format(i) for i in range(100)],
)  # fmt:skip

col1, col2, col3 = gui.columns([3, 1, 1])
with col1:
    desc_cols = ["shortname", "shorttype", "shortpop", "shortage", "unit"]
    format_variable = lambda v: (
        " | ".join(map(str, meta[meta["variable"] == v][desc_cols].values[0]))
    )
    variable = gui.selectbox(
        "Select a variable",
        meta["variable"],
        format_func=format_variable,
    )
with col2:
    year = gui.selectbox("Select a year", sorted(df["year"].unique(), reverse=True))
with col3:
    group_name = gui.selectbox("Select a group", list(groups.keys()))


# filter data
df = df[df["variable"] == variable]
df = df[df["year"] == year]
df = df[df["percentile"].apply(lambda x: x in groups[group_name])]

if not len(df):
    gui.error("No data available")
    gui.stop()

# drop unused columns
df = df[["percentile", "value"]]

# convert percentile to ranges
df["percentile"] = df["percentile"].apply(parse_percentile, convert_dtype=True)
gui.write(df.sort_values(by="value"), hide_index=True)
df["percentile"] = df["percentile"].apply(lambda x: x[0])

df = df.sort_values(by="percentile")

my_records = []
for cutoff in [10, 50, 90, 95, 99, 99.9, 99.99, 99.999]:
    value = int(df[df["percentile"] == cutoff]["value"].values[0])
    inr = format_currency(value, "INR", locale="en_IN")
    my_records.append({"percentile": cutoff, "value": inr})
gui.dataframe(pd.DataFrame.from_records(my_records), hide_index=True)

col1, col2 = gui.columns(2)
with col1:
    start = gui.number_input(
        "Start", min_value=0.0, max_value=100.0, step=0.001, value=0.0, format="%.3f"
    )
with col2:
    end = gui.number_input(
        "End", min_value=0.0, max_value=100.0, step=0.001, value=100.0, format="%.3f"
    )
df = df[df["percentile"].between(start, end)]

YAXIS_TYPE = gui.selectbox("Select a y-axis type", ["linear", "log"])

fig = go.Figure(
    data=[
        go.Scatter(
            x=df["percentile"],
            y=df["value"],
        ),
        go.Bar(
            x=df["percentile"],
            y=df["value"],
        ),
    ]
)
fig.update_layout(
    title_text="{} in {}".format(format_variable(variable), year),
    xaxis_title_text="Percentile",
    yaxis_title_text="Value",
    # log scale for y axis
    yaxis_type=YAXIS_TYPE,
    # show y axis as currency
    yaxis_tickformat=",.0f",
    # add a range slider and autscale the y-axis
    xaxis=dict(rangeslider=dict(visible=True)),
    yaxis=dict(autorange=True, fixedrange=False),
)

gui.plotly_chart(fig, use_container_width=True)