Top Performing Show HN Posts

from sturdystats import Index
from plotly import express as px
index = Index(id="index_f2ffe2f7901845c59c15aab45685fa3c")

## Plotting Settings
import copy
from plotly import io as pio
from plotly import express as px
px.defaults.template = "simple_white"  # Change the template
px.defaults.color_discrete_sequence = px.colors.qualitative.Dark24 # Change color sequence
pio.templates["no_margins"] = copy.deepcopy(pio.templates["simple_white"])
pio.templates["no_margins"].layout.margin = {"l": 0, "r": 0, "t": 30, "b": 0, "pad": 0}
pio.templates["no_margins"].layout.plot_bgcolor = 'rgba(0,0,0,0)'
pio.templates["no_margins"].layout.paper_bgcolor = 'rgba(0,0,0,0)'
def procFig(fig):
    fig.update_layout(
        font_family="Charter",
        autosize=True,
        plot_bgcolor= "rgba(0, 0, 0, 0)",
        paper_bgcolor= "rgba(0, 0, 0, 0)",
        margin=dict(
            l=0,
            r=0,
            b=0,
            t=30,
            pad=0
        ),
        coloraxis_colorbar=dict(
        orientation='h',  # horizontal orientation
        yanchor='top',
        y=-0.1,  # position below the plot (negative values go below)
        xanchor='center',
        x=0.5,  # center horizontally
        thickness=15,  # height of the colorbar
        len=0.75  # length as fraction of plot width
    )
    )
    fig.layout.xaxis.fixedrange = True
    fig.layout.yaxis.fixedrange = True
    return fig
    
pio.templates.default = "no_margins"

Explore Topics

topic_df = index.topicSearch(limit=500)
topic_df.head(5)

	short_title	topic_id	mentions	prevalence	topic_group_id	topic_group_short_title	conc	entropy
0	Personal Projects	60	20745.0	0.101177	3	Projects and Personal Development	90.538170	6.966460
1	Open-Source Hosting	145	16622.0	0.039764	5	Cloud and Infrastructure	18.127928	6.122626
2	Scripting and Automation	104	13892.0	0.038917	6	Tools and Utilities	17.364424	5.922157
3	Web Development Tools	141	11666.0	0.027827	0	Web and Software Development	18.536110	5.498559
4	Social Photo Sharing	135	10991.0	0.026396	4	Interactive Media and Games	18.562660	5.140318

topic_df["title"] = "Show HN"
fig = px.sunburst(
    topic_df,
    path=["title", "topic_group_short_title", "short_title"],
    values="prevalence",
    height=500
)
fig = procFig(fig)
fig

Relationship between Topics & Points

df = index.queryMeta("""
WITH t AS (
    SELECT 
        UNNEST(sum_topic_counts_vals) as topic_vals,
        UNNEST(sum_topic_counts_inds) - 1 as topic_id, -- duckdb is 1 indexed
        year(published::DATE) as year,
        score
    FROM doc
)
SELECT 
    topic_id,
    year,
    count(*) as n_posts,
    avg(score) as avg_score,
    avg((score > 10)::int) as p10,
    avg((score > 100)::int) as p100,
FROM t
WHERE topic_vals > 5
GROUP BY topic_id, year
ORDER BY topic_id, year
""", paginate=True)
df.sample(5)

	topic_id	year	n_posts	avg_score	p10	p100
1177	92	2013	10	6.600000	0.200000	0.000000
1872	150	2018	15	7.066667	0.133333	0.000000
662	52	2021	5	3.000000	0.000000	0.000000
975	77	2017	10	26.000000	0.200000	0.100000
1471	115	2025	1131	13.419982	0.129973	0.032714

Add Topic Information

tmp = topic_df.set_index("topic_id").to_dict()

df["topic"] = (df.topic_id).apply(tmp["short_title"].get)
df["topic_group"] = (df.topic_id).apply(tmp["topic_group_short_title"].get)
df = df.dropna().copy()
df.sample(5)

	topic_id	year	n_posts	avg_score	p10	p100	topic	topic_group
129	10	2018	33	18.121212	0.181818	0.060606	DIY Hardware IoT Projects	Projects and Personal Development
53	5	2011	1	48.000000	1.000000	0.000000	AI-Driven Personalization	Artificial Intelligence
1817	146	2021	13	32.307692	0.307692	0.076923	Wikipedia Article Tracking	Data and Analytics
618	49	2018	8	5.875000	0.125000	0.000000	Creative Computing Commands	Interactive Media and Games
2053	167	2014	83	8.120482	0.168675	0.012048	Crypto Commerce	Finance and Commerce

Add a Prior

We don’t want to overindex on topics that have only a few submissions per year but do extremely well. To add some smoothening, we add a simple prior that assumes 2/50 posts receive 100 points. This penalized topics that only have a few submissions.

df["P(score>100)"] = df.apply(lambda x: (x["p100"]*x["n_posts"]+ 2)/ (x["n_posts"]+50), axis=1)

Let’s explore 2025

Top Ranking Topics

tmp = df.loc[df.year==2025]
tmp.sort_values("P(score>100)", ascending=False).head()[["topic", "P(score>100)"]]

	topic	P(score>100)
136	DIY Hardware IoT Projects	0.093923
1031	Open Source Projects	0.087985
2132	Error Handling and Debugging	0.078652
1997	Programming Language Interpreters	0.077095
1458	Life Narratives	0.067568

Visual Overview

df["title"] = "Show HN<br> Performance<br>2025"
fig = px.sunburst(df.loc[df.year==2025], path=["title", "topic_group", "topic"], 
            values="n_posts", color_continuous_scale="greens",
            color="P(score>100)", height=600)
procFig(fig)

Comparing across years

Treemap: Click around to explore!

df["title"] = "Show HN<br>Performance"
fig = px.treemap(df, path=["title", "year", "topic_group", "topic"], 
            values="n_posts", color_continuous_scale="greens",
            color="P(score>100)", height=600,  )
procFig(fig)

2022-2024 stand out

It looks like Show HN really start performing amazingly well during this stretch, with a return to normalcy in 2025.

Unlock Your Unstructured Data Today

from sturdystats import Index

index = Index("Custom Analysis")
index.upload(df.to_dict("records"))
index.commit()
index.train()

# Ready to Explore 
index.topicSearch()

More Examples