streamlit.py

from collections import namedtuple
import altair as alt
import math
import pandas as pd
import streamlit as st
import polars as pl
import openmeteo_requests
from datetime import *

st.set_page_config(
    page_title="Weather Data",
    page_icon="🌤️",
	layout='wide'
)

st.markdown("# Historic Weather Data for 3 BYU Campuses")
st.markdown("__Explore weather data gathered at all 3 BYU locations from any chosen date range. Observe visuals explaining the temperature, pressure, wind speed, and far more from all locations.__")

columns = st.columns(3, gap='medium')

openmeteo = openmeteo_requests.Client()

st.sidebar.markdown("_Note: Data is only present up to 4 days ago_")

length = st.sidebar.slider("Length of Time Period for Data (Days)", min_value=5, max_value=30, value=15)

campus = ['provo', 'hawaii', 'idaho']
locations = [(40.2518, 111.6493), (21.6419, 157.9267), (43.8145, 111.7833)]

start_day = st.sidebar.date_input("Start of Period for Data", value=(datetime.today() - timedelta(days=35)), min_value=date(2020, 1, 1), max_value=(datetime.today() - timedelta(days=(length+4))))

end = (start_day + timedelta(days=length)).strftime('%Y-%m-%d')
start = start_day.strftime('%Y-%m-%d')

zone_select = st.sidebar.selectbox("Choose Timezone", options=['Mountain Time', 'Hawaiian Time'])

zone = 'HST' if zone_select == 'Hawaiian Time' else 'MST'

variables = ['Temperature (F)', 'Apparent Temp (F)', 'Humidity (%)', 'Pressure (hPa)', 'Precipitation (in)', 'Rain (in)', 'Snowfall (in)', 'Cloud Cover (%)', 'Wind Speed (mph)', 'Snow Depth (m)']
variable_select = st.sidebar.selectbox("Choose variable of interest", options=variables)
variable_dict = {
	'Temperature (F)': 'temperature_2m',
	'Apparent Temp (F)': 'apparent_temperature',
	'Humidity (%)': 'relative_humidity_2m',
	'Pressure (hPa)': 'surface_pressure',
	'Precipitation (in)': 'precipitation',
	'Rain (in)': 'rain',
	'Snowfall (in)': 'snowfall',
	'Cloud Cover (%)': 'cloud_cover',
	'Wind Speed (mph)': 'wind_speed_10m',
	'Snow Depth (m)': 'snow_depth'
}

var = variable_dict[variable_select]

hi_lo = st.sidebar.selectbox("Daily Extremes", options=['Max', 'Min'])

daily_var = 'temperature_2m_max' if hi_lo == 'Max' else 'temperature_2m_min'

def request(url, lat, long, start, end, variable, time_unit, timezone):
    params = {
	"latitude": lat,
	"longitude": long,
	"start_date": start,
	"end_date": end,
	time_unit: variable,
    "temperature_unit": "fahrenheit",
	'timezone': timezone,
	'precipitation_unit': 'inch',
	'wind_speed_unit': 'mph'
	}
    return openmeteo.weather_api(url, params=params)[0]

i = 0
temp_box = pl.DataFrame()
for location in locations:
	actual = request("https://archive-api.open-meteo.com/v1/archive", location[0], location[1], start, end, var, 'hourly', zone)
	forecasted = request("https://historical-forecast-api.open-meteo.com/v1/forecast", location[0], location[1], start, end, var, 'hourly', zone)

	hourly_actual= actual.Hourly()
	hourly_var_actual = hourly_actual.Variables(0).ValuesAsNumpy()

	hourly_forecasted = forecasted.Hourly()
	hourly_var_forecasted = hourly_forecasted.Variables(0).ValuesAsNumpy()

	hourly_data = {"date-time": pd.date_range(
		start = pd.to_datetime(hourly_actual.Time(), unit = "s", utc = True),
		end = pd.to_datetime(hourly_actual.TimeEnd(), unit = "s", utc = True),
		freq = pd.Timedelta(seconds = hourly_actual.Interval()),
		inclusive = "left"
	)}

	hourly_data[var] = hourly_var_actual
	hourly_data[f"forecast_{var}"] = hourly_var_forecasted

	hourly_dataframe = pl.DataFrame(hourly_data)

	hourly_dataframe = hourly_dataframe.with_columns(pl.col('date-time').dt.strftime('%Y-%m-%d').alias('date'))
	hourly_dataframe = hourly_dataframe.with_columns(pl.col('date-time').dt.strftime("%H:%M:%S").alias('time'))

	with columns[i]:
		st.markdown(f"### {campus[i].title()}")
		st.dataframe(hourly_dataframe.drop('date-time'))
		st.dataframe(hourly_dataframe.select([var, f'forecast_{var}']).describe()[2:9])
      
	current_campus = hourly_dataframe.with_columns(pl.lit(campus[i]).alias('campus'))

	end_month = (datetime.today() - timedelta(days=4)).strftime('%Y-%m-%d')
	start_month = (datetime.today() - timedelta(days=35)).strftime('%Y-%m-%d')
	daily = request("https://archive-api.open-meteo.com/v1/archive", location[0], location[1], start_month, end_month, daily_var, 'daily', zone).Daily()

	daily_data = {"date-time": pd.date_range(
		start = pd.to_datetime(daily.Time(), unit = "s", utc = True),
		end = pd.to_datetime(daily.TimeEnd(), unit = "s", utc = True),
		freq = pd.Timedelta(seconds = daily.Interval()),
		inclusive = "left"
	)}
	daily_data['temperature'] = daily.Variables(0).ValuesAsNumpy()

	daily_df = pl.DataFrame(daily_data)
	daily_df = daily_df.with_columns(pl.lit(campus[i]).alias('campus'))

	if len(temp_box) == 0:
		temp_box = pl.DataFrame(current_campus)
		temp_line = daily_df
	else:
		temp_box = temp_box.vstack(current_campus)
		temp_line = temp_line.vstack(daily_df)

	i+=1

daily = alt.Chart(temp_line).mark_line().encode(
	alt.X("date-time"),
	alt.Y('temperature'),
	color='campus'
).properties(
    title='Daily Temp Extremes Over Last Month',
	width=300,
	height=300
)


hourly_box = alt.Chart(temp_box).mark_boxplot().encode(
	alt.X(var).scale(zero=False),
	alt.Y("campus")
).properties(
    title='Hourly Readings',
	width=400,
	height=300
)


grouped = temp_box.select(['time', var, 'campus']).group_by(['time', 'campus']).mean()

hourly_average = alt.Chart(grouped).mark_line().encode(
	x="time",
	y=var,
	color="campus"
).properties(
    title='Average Reading Given Time of Day',
	width=400,
	height=300
)


kpi_max = temp_box.select([var, 'campus']).group_by('campus').max()

pr_hi = kpi_max.filter(pl.col('campus') == 'provo')[var][0]
id_hi = kpi_max.filter(pl.col('campus') == 'idaho')[var][0]
hi_hi = kpi_max.filter(pl.col('campus') == 'hawaii')[var][0]


kpi_low = temp_box.select([var, 'campus']).group_by('campus').min()
pr_lo = kpi_low.filter(pl.col('campus') == 'provo')[var][0]
id_lo = kpi_low.filter(pl.col('campus') == 'idaho')[var][0]
hi_lo = kpi_low.filter(pl.col('campus') == 'hawaii')[var][0]


st.sidebar.altair_chart(daily)

columns[1].altair_chart(hourly_box)

columns[0].altair_chart(hourly_average)

with columns [2]:
	sub_cols = st.columns(2)
	with sub_cols[0]:
		st.markdown(f"Provo Max: {round(pr_hi, 1)}")
		st.markdown(f"Hawaii Max: {round(hi_hi, 1)}")
		st.markdown(f"Idaho Max: {round(id_hi, 1)}")
	
	with sub_cols[1]:
		st.markdown(f"Provo Min: {round(pr_lo, 1)}")
		st.markdown(f"Hawaii Min: {round(hi_lo, 1)}")
		st.markdown(f"Idaho Min: {round(id_lo, 1)}")