As part of our efforts to build optimized, intuitive keyboard layouts for Indian languages, we’ve conducted detailed character frequency analysis for Malayalam. This post presents our findings—showing which characters, vowels, and consonants occur most frequently in real-world usage based on large Malayalam dataset—and briefly discusses how this data influenced the keyboard design.

For an explanation of our overall keyboard design methodology, including the rationale behind layout decisions, please refer to our Tamil keyboard design post: Designing a New Input Method for Tamil

Dataset Used for Frequency Analysis

To ensure broad linguistic coverage and reliability, we used Malayalam corpora by combining text data from multiple sources, including:

• IndicCorp
• Swathantra Malayalam corpus
• Wikipedia Malayalam corpus
• News articles, blogs, etc.

Here is the high-level statistics of the corpora used in our analysis:

• Total words: 1.699 billion
• Unique words: 25.345 million

This large and diverse dataset gives us a realistic view of character usage in modern written Malayalam.

Overall Character Frequency Heatmap

Characters such as അ (a), മ (ma), ന (na), and ക (ka) appear with consistently dark shades, across their rows, indicating they are among the most frequently used. These are foundational phonemes in Malayalam and are common across both spoken and written forms. Characters like പ (pa), ത (tha), and ല (la) show moderate frequency across their respective row, suggesting they are contextually important but not as dominant. Their usage may vary depending on the domain (e.g., literary vs. conversational text).

Vowel Frequency Heatmap

To gain deeper insights, we examined the frequency of Malayalam vowels both in their standalone and combined consonant-vowel forms. This analysis was derived by summing the columns of our overall character frequency chart, offering a clear view of vowel usage patterns. The results are visualized in the chart below.

Malayalam has a rich vowel system, but a handful of vowels—especially അ, ഇ, ഉ, and എ—occur most frequently. Further as expected, the shorter vowels are more frequent than their longer versions. Note that these counts include the usage frequency of vowels as well as the vowel sign glyphs in their consonant vowel forms. The last cell in the heat map actually refers to the frequency of the Chandrakala (Virama) character.

Consonant Frequency Heatmap

In addition to vowels, we also analyzed the frequency of Malayalam consonants. This was done by summing the rows of the overall character frequency chart, which highlights how often each consonant appears across different vowel combinations. The resulting data provides a clearer picture of consonant usage patterns in the language.

Among consonants, we can see a clear dominance of ക്, ന്, ത് യ്, and ര്. These findings informed our decisions in distributing consonants across the keyboard layout for minimizing the finger movements during typing. In line with the InScript keyboard layout convention, each aspirated consonant was assigned to a specific key, while its corresponding unaspirated counterpart was placed in the corresponding shift position.

Chillaksharam Frequency Heatmap

Impact on Keyboard Layout Design

Whether you're using a Malayalam phonetic keyboard or a standard Malayalam keyboard, these insights can guide better input method development. As part of our design, we excluded individual vowel signs and instead generate them dynamically from consonant-vowel combinations. This approach reduces complexity and prevents invalid character sequences.

These frequency trends helps in designing the layout for our Malayalam keyboard. High-frequency characters were assigned to primary home row or easy index finger positions, ensuring reduced typing effort and faster input for users. Low-frequency characters were positioned in secondary or long-press locations. As part of our design, we excluded individual vowel signs and instead generate them dynamically from consonant-vowel combinations. This approach reduces complexity and prevents invalid character sequences.

Explore Frequency Analyses in Other Languages

We’ve performed similar analysis for other Indian languages as well. Explore them below:

• Character Frequency Analysis for Hindi
• Character Frequency Analysis for Kannada
• Character Frequency Analysis for Tamil (includes design principles)
• Character Frequency Analysis for Telugu