Datawarehouse? No, data kitchen! Get the right ingredients and serve on a plate!
Joined May 2019
- Tweets 147
- Following 55
- Followers 4
- Likes 309
26 Photos and videos
30 Jan 2023
Today's bets after lunch. More in the green, but heavier in the red.
1
31
Datapancake retweeted
30 Sep 2022
@felipehoffa is demonstrating that your cloud database is not more… a database. I foresee that data lineage, data quality, data observability, ML & data science engine and data ingestion can be one product hoffa.medium.com/facebook-pr…
18
29
7 Sep 2020
Sweden, ages 15-29, deaths with Covid-19 2020 vs deaths by drug overdose 1997-2019.
3 Sep 2020
Scikit like Data prep in SQL is all cool, AI models are harder.
I tried by building an image recognition CNN in SnowflakeDB. Got as far as the first few steps before deciding ”naaa...!”
Looping, recursion just doesn’t work. Has to be in mem dataframe ops.
1
23 Aug 2020
Watched the CL-final. Half of the excitement gone with no crowds. Football’s magic lies in the connect between pitch and terrace. Take away one and not much remain.
15 Aug 2020
Bought two bottles of Champagne for home delivery at a slightly higher price from a local wine merchant rather than from Amazon. We need to keep local business going.
13 Aug 2020
When you tweet, how will you be labeled?
Analysis for topic ”Pharma”, a few days Nov 2019.
6 Aug 2020
Performance testing row level security cake bake in the data kitchen:
1. python (rando user generation and mapping to ranges of dim values)
2. node.js (random sessions per user, executing asynch queries on fact joind w user-dim-mapping)
3. @SnowflakeDB to execute on massive scale
1
Datapancake retweeted
30 Jul 2020
State of California COVID-19 Data is Now Available to Public via @SnowflakeDB Data Marketplace snowflake.com/news/state-of-… #COVID19 ❄️
3
3
28 Jul 2020
Brings so much hope!
28 Jul 2020
Breaking News: A long-awaited blood test for Alzheimer’s is in reach, scientists say. It could speed up treatment research and aid in diagnosing dementia patients.
nyti.ms/3jN5V9Q
24 Jul 2020
I mistakenly started to post fact based replies, with statistical significance accounted for, in a BBC Covid thread. Remembered the meaning of ”futile” and went back to delete them all. Wasted 2 hours of good life, still saved another 20 at risk
10 Jul 2020
Fast food is the new terrorism
10 Jul 2020
“They saw two people in there and they were ordering 20-odd meals at 1:30 this morning." A big KFC order helped tip off the Australian authorities that 16 people had broken coronavirus restrictions by attending a surprise birthday in a suburb of Melbourne. nyti.ms/3iWJazZ
8 Jul 2020
I’m currently trying out @StarschemaLtd ’s shared Covid dataset on @SnowflakeDB which has a new analytic UI natively available.
Here’s a graph of some countries’ DailyDeaths/million numbers as 21d moving avg.
7 Jul 2020
Converting Teradata to Snowflake SQL is truly data mining. Like hammer and chisel 19th century Dickensian coal data mining. Carrying a canary in a cage.
6 Jul 2020
Building AS OF JOINS for TS-analysis in #SnowflakeDB My solution uses LAG/LEAD to persist valid period per record, to join on. Good performance so far. Meaning low single seconds over 1Brows x 1Mrows
1
6 Jul 2020
Actually 1Brows x 100Mrows (one year of sensor-A data @ 100/s as of joined one year of sensor-B @ 10/s). And I’m running on a low-end size cluster.