Snowflake 마켓플레이스를 통한 서드 파티 데이터에 대한 직접 액세스

개요

사전 필요 조건 및 지식

알아볼 내용

구축할 것

워크시트 생성 및 SQL 복사

개요

1단계 - URL을 통해 Snowflake에 액세스

2단계 - Snowflake에 로그인

3단계 - Worksheets로 이동

4단계 - 워크시트 생성

5단계 - 워크시트 이름 바꾸기

6단계 - GitHub에서 Quickstart SQL에 액세스

7단계 - GitHub에서 설정 SQL 복사

8단계 - 설정 SQL을 GitHub에서 Snowflake 워크시트로 붙여넣기

9단계 - 다음 클릭 -->

자사 데이터에서 매출이 0인 날짜 조사

개요

1단계 - 추세를 위해 판매 시점 데이터 쿼리

USE ROLE tasty_data_engineer;
USE WAREHOUSE tasty_de_wh;

SELECT 
    o.date,
    SUM(o.price) AS daily_sales
FROM frostbyte_tasty_bytes.analytics.orders_v o
WHERE 1=1
    AND o.country = 'Germany'
    AND o.primary_city = 'Hamburg'
    AND DATE(o.order_ts) BETWEEN '2022-02-10' AND '2022-02-28'
GROUP BY o.date
ORDER BY o.date ASC;

2단계 - 다음 클릭 -->

Snowflake 마켓플레이스의 Weather Source 데이터 활용

개요

1단계 - Weather Source LLC: frostbyte Snowflake 마켓플레이스 목록 획득

2단계 - 자사 및 서드 파티 데이터 조정

CREATE OR REPLACE VIEW frostbyte_tasty_bytes.harmonized.daily_weather_v
    AS
SELECT 
    hd.*,
    TO_VARCHAR(hd.date_valid_std, 'YYYY-MM') AS yyyy_mm,
    pc.city_name AS city,
    c.country AS country_desc
FROM frostbyte_weathersource.onpoint_id.history_day hd
JOIN frostbyte_weathersource.onpoint_id.postal_codes pc
    ON pc.postal_code = hd.postal_code
    AND pc.country = hd.country
JOIN frostbyte_tasty_bytes.raw_pos.country c
    ON c.iso_country = hd.country
    AND c.city = hd.city_name;

3단계 - 일일 기온 시각화

SELECT 
    dw.country_desc,
    dw.city_name,
    dw.date_valid_std,
    AVG(dw.avg_temperature_air_2m_f) AS avg_temperature_air_2m_f
FROM frostbyte_tasty_bytes.harmonized.daily_weather_v dw
WHERE 1=1
    AND dw.country_desc = 'Germany'
    AND dw.city_name = 'Hamburg'
    AND YEAR(date_valid_std) = '2022'
    AND MONTH(date_valid_std) = '2'
GROUP BY dw.country_desc, dw.city_name, dw.date_valid_std
ORDER BY dw.date_valid_std DESC;

4단계 - 바람 데이터 가져오기

SELECT 
    dw.country_desc,
    dw.city_name,
    dw.date_valid_std,
    MAX(dw.max_wind_speed_100m_mph) AS max_wind_speed_100m_mph
FROM frostbyte_tasty_bytes.harmonized.daily_weather_v dw
WHERE 1=1
    AND dw.country_desc IN ('Germany')
    AND dw.city_name = 'Hamburg'
    AND YEAR(date_valid_std) = '2022'
    AND MONTH(date_valid_std) = '2'
GROUP BY dw.country_desc, dw.city_name, dw.date_valid_std
ORDER BY dw.date_valid_std DESC;

5단계 - 다음 클릭 -->

데이터 인사이트 민주화

개요

1단계- SQL 함수 생성

CREATE OR REPLACE FUNCTION frostbyte_tasty_bytes.analytics.fahrenheit_to_celsius(temp_f NUMBER(35,4))
RETURNS NUMBER(35,4)
AS
$$
    (temp_f - 32) * (5/9)
$$;

CREATE OR REPLACE FUNCTION frostbyte_tasty_bytes.analytics.inch_to_millimeter(inch NUMBER(35,4))
RETURNS NUMBER(35,4)
    AS
$$
    inch * 25.4
$$;

2단계 - 뷰를 위한 SQL 생성

SELECT 
    fd.date_valid_std AS date,
    fd.city_name,
    fd.country_desc,
    ZEROIFNULL(SUM(odv.price)) AS daily_sales,
    ROUND(AVG(fd.avg_temperature_air_2m_f),2) AS avg_temperature_fahrenheit,
    ROUND(AVG(frostbyte_tasty_bytes.analytics.fahrenheit_to_celsius(fd.avg_temperature_air_2m_f)),2) AS avg_temperature_celsius,
    ROUND(AVG(fd.tot_precipitation_in),2) AS avg_precipitation_inches,
    ROUND(AVG(frostbyte_tasty_bytes.analytics.inch_to_millimeter(fd.tot_precipitation_in)),2) AS avg_precipitation_millimeters,
    MAX(fd.max_wind_speed_100m_mph) AS max_wind_speed_100m_mph
FROM frostbyte_tasty_bytes.harmonized.daily_weather_v fd
LEFT JOIN frostbyte_tasty_bytes.harmonized.orders_v odv
    ON fd.date_valid_std = DATE(odv.order_ts)
    AND fd.city_name = odv.primary_city
    AND fd.country_desc = odv.country
WHERE 1=1
    AND fd.country_desc = 'Germany'
    AND fd.city = 'Hamburg'
    AND fd.yyyy_mm = '2022-02'
GROUP BY fd.date_valid_std, fd.city_name, fd.country_desc
ORDER BY fd.date_valid_std ASC;

3단계 - 분석 뷰 배포

CREATE OR REPLACE VIEW frostbyte_tasty_bytes.analytics.daily_city_metrics_v
COMMENT = 'Daily Weather Source Metrics and Orders Data for our Cities'
    AS
SELECT 
    fd.date_valid_std AS date,
    fd.city_name,
    fd.country_desc,
    ZEROIFNULL(SUM(odv.price)) AS daily_sales,
    ROUND(AVG(fd.avg_temperature_air_2m_f),2) AS avg_temperature_fahrenheit,
    ROUND(AVG(frostbyte_tasty_bytes.analytics.fahrenheit_to_celsius(fd.avg_temperature_air_2m_f)),2) AS avg_temperature_celsius,
    ROUND(AVG(fd.tot_precipitation_in),2) AS avg_precipitation_inches,
    ROUND(AVG(frostbyte_tasty_bytes.analytics.inch_to_millimeter(fd.tot_precipitation_in)),2) AS avg_precipitation_millimeters,
    MAX(fd.max_wind_speed_100m_mph) AS max_wind_speed_100m_mph
FROM frostbyte_tasty_bytes.harmonized.daily_weather_v fd
LEFT JOIN frostbyte_tasty_bytes.harmonized.orders_v odv
    ON fd.date_valid_std = DATE(odv.order_ts)
    AND fd.city_name = odv.primary_city
    AND fd.country_desc = odv.country
WHERE 1=1
GROUP BY fd.date_valid_std, fd.city_name, fd.country_desc;

4단계 - 다음 클릭 -->

매출 및 마켓플레이스 날씨 데이터에서 인사이트 도출

개요

1단계 - 분석 간소화

SELECT 
    dcm.date,
    dcm.city_name,
    dcm.country_desc,
    dcm.daily_sales,
    dcm.avg_temperature_fahrenheit,
    dcm.avg_temperature_celsius,
    dcm.avg_precipitation_inches,
    dcm.avg_precipitation_millimeters,
    dcm.max_wind_speed_100m_mph
FROM frostbyte_tasty_bytes.analytics.daily_city_metrics_v dcm
WHERE 1=1
    AND dcm.country_desc = 'Germany'
    AND dcm.city_name = 'Hamburg'
    AND dcm.date BETWEEN '2022-02-01' AND '2022-02-26'
ORDER BY date DESC;

2단계 - 다음 클릭 -->

종료 및 다음 단계

결론

다음 단계